Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The amendment filed 08/28/2026 has been entered. Claims 1 and 11 have been amended. Claim 10 has been canceled. Claims 1–9 and 11–20 are pending and are addressed in this Office action.
Response to Arguments
Applicant’s arguments filed 08/28/2026 have been fully considered but are not persuasive for the reasons set forth below.
To the extent Applicant argues that the applied references fail to teach the limitations added by amendment, namely (i) that “the plurality of road actors are different from the vehicle,” (ii) that “the driving style is indicative of dangerous or reckless driving,” and (iii) “automatically control[ling] the vehicle to at least one of” the recited maneuvers, the arguments are not persuasive. The arguments are directed to the claims as amended, and the rejection set forth below has been updated to address each amended limitation.
Regarding (i), Zhu (US 2012/0083960 A1) tracks “objects external to the vehicle such as other vehicles” (Zhu, [0032]), for example a tractor-trailer driving alongside vehicle 101 within its sensor fields (Zhu, [0055]). None of Zhu’s tracked road actors is the host vehicle itself.
Regarding (ii), Zhu expressly identifies other vehicles “behaving dangerously, for example, a swerving or speeding vehicle or distracted a driver talking on a mobile phone” (Zhu, [0109]). Kwon (US 2024/0239350 A1) infers the driving characteristic of an individual target vehicle from its driving data (Kwon, [0064]), a characteristic that distinguishes vehicles that are “aggressively or stably driven depending on a style of a driver” (Kwon, [0003]). The combination therefore determines a driving style of the first road actor that is indicative of dangerous or reckless driving.
Regarding (iii), Zhu’s computer itself actuates acceleration, braking, and steering in autonomous mode (Zhu, [0027]). In response to the characterized behavior of another vehicle, it slows the vehicle (Zhu, [0051]), changes lanes or stays a particular distance away (Zhu, [0059]), and keeps its passengers “away from” vehicles behaving dangerously (Zhu, [0109]). Because the claim recites the maneuvers in the alternative, any one of them satisfies the limitation (see Claim Interpretation, below). Kwon likewise teaches that the inferred driving characteristic of the target vehicle “may be used to control the learning driving vehicle” (Kwon, [0071]).
To the extent the claims were previously rejected under 35 U.S.C. 101, that rejection is withdrawn. Although the claims recite steps of evaluating relative kinematic parameters and determining a driving style, which are mental processes, amended claims 1 and 11 additionally require, based on the determined driving style, automatically controlling the vehicle to perform at least one of a set of specific physical maneuvers. This limitation applies the judicial exception in a meaningful way beyond generally linking it to a technological environment, and integrates the exception into a practical application. See MPEP 2106.04(d) and 2106.05(e).
Claim Objections
The claims are objected to because of the following informalities:
Claims 1 and 11: “automatically control[ling] the vehicle to at least one of reduce a speed of the vehicle, change a lane of the vehicle, pull the vehicle onto a road shoulder or a parking area, increase the speed of the vehicle, overtake the first road actor, or increase a distance between the vehicle and the first road actor” is grammatically incorrect. It is suggested that the limitation read: “automatically control[ling] the vehicle to perform at least one of: reducing a speed of the vehicle, changing a lane of the vehicle, pulling the vehicle onto a road shoulder or a parking area, increasing the speed of the vehicle, overtaking the first road actor, or increasing a distance between the vehicle and the first road actor.”
Claims 3 and 13: “the one or more relative kinematic parameters” should read “the relative kinematic parameters” for consistency with claims 1 and 11, which do not recite “one or more” relative kinematic parameters.
Claims 4, 9, 11, 16, and 19: “the first actor” should read “the first road actor” for consistency with the terminology used throughout the claims.
Claim 12: “a yaw tuning rate” should read “a yaw turning rate” (compare claim 2).
Claim 18: “the road first actor” should read “the first road actor.”
Claim 20: “at least one of”” should read “at least one of:” (the closing quotation mark appears to be a typographical error).
Appropriate correction is required.
Claim Rejections – 35 U.S.C. § 112
35 U.S.C. 112(a) – Written Description (New Matter)
The following is a quotation of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
Claims 1–9 and 11–20 are rejected under 35 U.S.C. 112(a) as failing to comply with the written description requirement. The claims contain subject matter that was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, at the time the application was filed, had possession of the claimed invention. The subject matter at issue was introduced by the amendment filed 08/28/2026 and constitutes new matter. See MPEP 2163.06.
The limitations at issue. As amended, independent claim 1 requires “determin[ing] ... a driving style of the first road actor, wherein the driving style is indicative of dangerous or reckless driving” and, “based on the driving style of the first road actor, automatically control[ling] the vehicle to at least one of ... increase the speed of the vehicle, overtake the first road actor, ...” Independent claim 11 recites the same limitations in method form. Because the control limitation is recited in the alternative, each alternative defines a separate embodiment within the scope of the claims. Two of those embodiments are an ego vehicle that, because it has determined that the first road actor is driving dangerously or recklessly, automatically (a) increases its speed or (b) overtakes that road actor. The original disclosure must show possession of each claimed alternative. MPEP 2163(II)(A)(3)(a).
The specification describes increasing speed and overtaking only for a slow road actor, not a dangerous one. The only disclosure of increasing speed or overtaking in response to a determined driving style is in [0073]:
“if road actor 404a that is in front or ahead of the ego vehicle 100 is determined to be driving very slow, the autonomy computing system 200 and/or the behaviors and planning module 238 can determine to cause the ego vehicle to increase speed and overtake the road actor 404a.”
Driving “very slow” is not the dangerous or reckless driving style now recited in the claims. The same paragraph separately describes the response to a road actor with a dangerous driving style, and that response is neither increasing speed nor overtaking:
“If a road actor 404 is determined to be exhibiting a dangerous driving style, the autonomy computing system 200 and/or the behaviors and planning module 238 can cause the ego vehicle 100 to pull into a road shoulder, if safe or possible, or into a parking area. When pulling into the road shoulder or parking place, the autonomy computing system 200 and/or the behaviors and planning module 238 can cause the ego vehicle 100 to navigate in a manner that causes the ego vehicle 100 to stay as far as possible from the road actor 404 exhibiting the dangerous driving style.” ([0073])
The specification consistently teaches moving away from a dangerous or reckless road actor. Every passage addressing such a road actor describes avoiding it or moving away from it:
“One technical challenge faced by ADSs is how to detect road actors with unsafe, risky or dangerous driving styles, and determine the proper action(s) to be taken to avoid or stay away from such road actors.” ([0024])
“the autonomy computing system 200 and/or the behaviors and planning module 238 can take proper actions or decisions to keep the ego vehicle 100 away from any road actor 404 exhibiting dangerous or reckless driving style.” ([0050])
Overtaking a road actor requires the ego vehicle to close on that road actor, travel alongside it, and pass it, and increasing speed is the means by which the ego vehicle closes on a road actor ahead. Both are the opposite of the “avoid or stay away” response that the specification consistently associates with a dangerous or reckless driving style. Nothing in the original disclosure reasonably conveys that the inventors had possession of a system that responds to a dangerous or reckless road actor by increasing speed or by overtaking it.
The examiner has considered that increasing speed could, in some circumstances, increase the separation from a road actor located behind the ego vehicle. The specification, however, never describes that response. Its only example of responding to a road actor behind the ego vehicle is changing lanes away from a tailgater, and its only example of increasing speed is overtaking a slow road actor ahead (both at [0073]). Whether increasing speed would have been a sensible response to a dangerous road actor is a question of obviousness, not possession, and cannot supply written description support, as explained below.
Paragraph [0072] does not supply the missing support. Paragraph [0072] states that the actions “can include at least one of reducing vehicle speed, changing lanes, pulling into a road shoulder, increasing speed, and/or overtaking a road actor ahead, among other possible actions.” That passage is a generic list of available maneuvers. It does not identify which maneuver is taken in response to which driving style. Paragraph [0073] supplies that correlation, and it pairs increasing speed and overtaking with a slow road actor and pairs a dangerous road actor with pulling over and staying as far away as possible.
The claimed combination can be assembled only by selecting “dangerous or reckless” from the list of driving styles in [0022] (one of more than ten listed styles) and “increasing speed” or “overtaking” from the separate list of maneuvers in [0072]. The specification provides no “blaze marks” directing one skilled in the art to either pairing. A specification that discloses lists of elements does not, without more, provide written description support for a specific combination drawn from those lists. See Purdue Pharma L.P. v. Faulding Inc., 230 F.3d 1320, 1326–28 (Fed. Cir. 2000); In re Ruschig, 379 F.2d 990, 994–95 (CCPA 1967).
Nor is it sufficient that the claimed pairings might be obvious from, or fall within the general teaching of, the disclosure. “[A] description that merely renders the invention obvious does not satisfy the requirement.” Ariad Pharms., Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1352 (Fed. Cir. 2010) (en banc); Lockwood v. Am. Airlines, Inc., 107 F.3d 1565, 1571–72 (Fed. Cir. 1997). The test is whether the disclosure, as originally filed, reasonably conveys possession of the subject matter now claimed. Ariad, 598 F.3d at 1351. For the reasons above, it does not.
Claims 2–9 and 12–20 depend directly or indirectly from claim 1 or claim 11. They do not remove the unsupported alternatives, so they are rejected for the same reason.
Applicant is advised that deleting the alternatives “increase the speed of the vehicle” and “overtake the first road actor” from claims 1 and 11 would overcome this rejection. Alternatively, Applicant may point to specific disclosure, by paragraph number, that supports the amended limitation. See MPEP 714.02 and 2163(II)(A).
35 U.S.C. 112(b) – Indefiniteness
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claim 20 is rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
Claim 20 recites the limitation “the one or more actions to be taken by the vehicle.” There is insufficient antecedent basis for this limitation in the claim. As originally filed, claim 11 recited “determining one or more actions to be taken by the vehicle based on the driving style of the first road actor,” and that recitation provided antecedent basis for this phrase in claim 20. The amendment filed 08/28/2026 deleted that recitation. Claim 11 now recites “automatically controlling the vehicle to at least one of reduce a speed of the vehicle, change a lane of the vehicle, pull the vehicle onto a road shoulder or a parking area, increase the speed of the vehicle, overtake the first road actor, or increase a distance between the vehicle and the first road actor.”
As a result, it is unclear whether “the one or more actions” in claim 20 refers to the maneuvers recited in the controlling step of claim 11 or to a different, undefined set of actions. That uncertainty also affects the rest of claim 20. If “the one or more actions” are the maneuvers of claim 11, it is unclear whether claim 20 limits the controlling step of claim 11 to the three recited maneuvers. If they are a separate set of actions, it is unclear whether, or how, those actions are performed at all, since claim 11 no longer recites determining or taking any “actions.” The scope of claim 20 therefore cannot be determined with reasonable certainty. Nautilus, Inc. v. Biosig Instruments, Inc., 572 U.S. 898, 901 (2014); MPEP 2173.05(e).
For purposes of examination, “the one or more actions to be taken by the vehicle” is interpreted as the at least one maneuver performed in the automatically controlling step of claim 11, selected from reducing vehicle speed, changing lanes, or pulling into a road shoulder based on a relative position of the first road actor with respect to the vehicle.
Applicant may overcome this rejection by, for example, amending claim 20 to recite: “wherein the automatically controlling the vehicle comprises selecting the at least one of reducing the speed of the vehicle, changing the lane of the vehicle, or pulling the vehicle onto the road shoulder based on a relative position of the first road actor with respect to the vehicle.” Alternatively, Applicant may cancel claim 20, consistent with the cancellation of its parallel system claim 10.
Applicant is further advised that claim 20 also recites “the relative position of the first road actor with respect to the vehicle,” whereas claim 11 recites relative kinematic parameters of the first road actor only with respect to other road actors and road features. The suggested amendment above recites “a relative position,” which would also resolve any question of antecedent basis for that phrase.
Claim Rejections – 35 U.S.C. § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
Determining the scope and contents of the prior art.
Ascertaining the differences between the prior art and the claims at issue.
Resolving the level of ordinary skill in the pertinent art.
Considering objective evidence present in the application indicating obviousness or nonobviousness.
Rejection I: Zhu in view of Kwon
Claims 1–4, 8, 11–14, 18, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Zhu et al. (US 2012/0083960 A1, hereinafter “Zhu”) in view of Kwon et al. (US 2024/0239350 A1, hereinafter “Kwon”).
Regarding Claim 1,
Disclosure by Zhu
Zhu discloses:
A system for monitoring road actors, the system comprising:
See at least:
“a vehicle implementing the autonomous driving system is capable of detecting and reacting to surrounding objects. Some of the detected objects will be mobile, such as pedestrians, automobiles, and bicycles.” (Zhu, [0017])
Rationale:
Zhu expressly discloses an autonomous driving system whose function is to detect, identify, and react to the mobile objects (automobiles, bicycles, pedestrians) around the vehicle. Those mobile objects are road actors, and continuously detecting and reacting to them is monitoring them. The open-ended transitional term “comprising” permits Zhu's additional components (e.g., navigation and display systems) without affecting the mapping.
one or more sensors of a vehicle
See at least:
“The vehicle may include components for detecting objects external to the vehicle such as other vehicles, obstacles in the roadway, traffic signals, signs, trees, etc. The detection system may include lasers, sonar, radar, cameras or any other detection devices.” (Zhu, [0032])
Rationale:
Zhu expressly discloses one or more sensors (lasers, sonar, radar, cameras) that are components of the vehicle itself, i.e., sensors of a vehicle.
to detect and track a plurality of road actors
See at least:
“The sensors described may be used to identify, track and predict the movements of pedestrians, bicycles, other vehicles, or objects in the roadway.” (Zhu, [0042])
Rationale:
Zhu expressly discloses that the sensors are used to detect (identify) and track the movements of multiple road users at once (pedestrians, bicycles, other vehicles), which is a plurality of road actors. Zhu's FIG. 6 flow applies this detection and tracking to each object entering the sensor fields (Blocks 620-670).
in a surrounding environment of the vehicle,
See at least:
“the sensors may provide the location and shape information of objects surrounding the vehicle to computer 110, which in turn may identify the object as another vehicle.” (Zhu, [0042])
Rationale:
Zhu expressly discloses that the tracked objects are those “surrounding the vehicle,” i.e., located in the surrounding environment of the vehicle.
wherein the plurality of road actors are different from the vehicle;
See at least:
“FIGS. 5A and 5B depict vehicle 101 as driving along side a tractor-trailer 510. As provided in FIG. 5B, the tractor-trailer 510 is within range of several sensor fields of vehicle 101.” (Zhu, [0055])
Rationale:
Zhu expressly discloses that the tracked road actor (tractor-trailer 510) is a separate vehicle driving alongside the host vehicle 101 and detected within the host vehicle's sensor fields. Zhu's tracked objects are, throughout, “objects external to the vehicle” ([0032]); none of them is vehicle 101 itself. The road actors are therefore different from the vehicle.
a processing device
See at least:
“The vehicle may have one or more computers, such as computer 110 containing a processor 120, memory 130 and other components typically present in general purpose computers.” (Zhu, [0018])
Rationale:
Zhu expressly discloses a processing device (processor 120 of computer 110).
in communication with the one or more sensors,
See at least:
“Many of these sensors provide data that is processed by the computer in real-time, that is, the sensors may continuously update their output to reflect the environment being sensed at or over a range of time, and continuously or as demanded provide that updated output to the computer” (Zhu, [0034])
Rationale:
Zhu expressly discloses that the sensors provide their output to the computer in real time. Providing that output requires a communication path between the sensors and the processor, so the processing device is in communication with the one or more sensors.
the processing device is configured to execute instructions stored in a memory
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses instructions 132 stored in memory 130 that are executed by processor 120, i.e., a processing device configured to execute instructions stored in a memory.
to perform operations comprising:
See at least:
“The instructions 132 may be any set of instructions to be executed directly (such as machine code) or indirectly (such as scripts) by the processor. ... Functions, methods and routines of the instructions are explained in more detail below.” (Zhu, [0020])
Rationale:
Zhu expressly discloses that the stored instructions carry out the functions, methods, and routines described in the remainder of the specification. Executing those instructions performs operations, and each of the operations mapped below is a routine of those instructions.
receive, from the one or more sensors, sensor data
See at least:
“Upon detecting the object, the vehicle's computer 110 may classify the object based on the data received by the vehicle's sensors (Block 630).” (Zhu, [0058])
Rationale:
Zhu expressly discloses that computer 110 receives data from the vehicle's sensors concerning the detected objects, i.e., sensor data received from the one or more sensors.
indicative of kinematic parameters of the plurality of road actors;
See at least:
“the computer system 110 may simply predict the object's future movement based solely on the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly discloses that the received sensor data yields each tracked object's direction, acceleration/deceleration, and velocity. These are kinematic parameters. Because Zhu tracks multiple objects ([0042]), the data is indicative of kinematic parameters of the plurality of road actors.
determine, using the sensor data, relative kinematic parameters of a first road actor of the plurality of road actors
See at least:
“any type of environmental or contextual information may be used to determine the object's current state, such as the object's speed, route the object has traveled, nature of the roadway on which the object is traveling, any lane changes made by the object, or the object's use of headlights or blinkers.” (Zhu, [0054])
Rationale:
Zhu expressly discloses determining, from the sensor data, a state of a particular tracked object (a first road actor of the plurality) that includes its motion in relation to its context: the route it has traveled, its lane changes, and the roadway it travels on. Motion expressed in relation to something other than the host vehicle is a relative kinematic parameter. What that motion is relative to is mapped separately below.
and relative to one or more road features;
See at least:
“data collected from the sensors of vehicle 101 may be used to determine that the tractor-trailer is traveling at 55 miles per hour in the left-hand lane of the interstate.” (Zhu, [0055])
Rationale:
Zhu expressly discloses determining the first road actor's motion relative to road features: its speed within a specific lane (the left-hand lane) of a specific roadway (the interstate). Zhu's stored map includes “lane lines” and “speed limits” for each roadway segment ([0052]), so the object's lane position and speed are determined relative to one or more road features.
determine, using the kinematic parameters and the relative kinematic parameters of the first road actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640). Upon determining the objects classification and state, computer 110 accesses behavior model information contained in database 137 (Block 650).” (Zhu, [0058])
Rationale:
Zhu expressly discloses using the object's speed (a kinematic parameter) and its lane position (a relative kinematic parameter with respect to a road feature) as the inputs from which the object's likely behavior is determined. Zhu does not call the output a “driving style”; that limitation is addressed separately below. In the combination, the relative kinematic parameters with respect to other road actors (taught by Kwon, below) are added to these same inputs.
wherein the driving style is indicative of dangerous or reckless driving; and
See at least:
“Autonomous vehicles may also keep passengers away from other humans or vehicles (autonomous or not) behaving dangerously, for example, a swerving or speeding vehicle or distracted a driver talking on a mobile phone.” (Zhu, [0109])
Rationale:
Zhu expressly discloses that the autonomous vehicle characterizes another vehicle's driving as dangerous, with examples (swerving, speeding, distracted driving) that are classic indicators of dangerous or reckless driving. Swerving is evident from the vehicle's lateral motion relative to its lane, and speeding from its velocity relative to the posted limit that Zhu stores for each roadway segment ([0052]). Once Zhu's system determines a driving style of the road actor (as taught by Kwon, below), a PHOSITA would have found it obvious for that style to be one indicative of dangerous or reckless driving, because identifying dangerous drivers is the express purpose for which Zhu characterizes surrounding vehicles in this passage.
based on the driving style of the first road actor,
See at least:
“The new control strategy is therefore specific to the predicted behavior of tractor-trailer 510, in that a different behavior model and a different control strategy would be used if object 510 was a motorcycle or sports car.” (Zhu, [0059])
Rationale:
Zhu expressly discloses selecting the vehicle's control response based on the characterization of the particular road actor's behavior, so that a different characterization produces a different control strategy. Combined with Zhu [0109] (the vehicle reacts to vehicles behaving dangerously) and Kwon's driving style (below), the control is performed based on the driving style of the first road actor.
automatically control the vehicle
See at least:
“if the vehicle is operating in a completely autonomous mode, computer 110 may cause the vehicle to accelerate (e.g., by increasing fuel or other energy provided to the engine), decelerate (e.g., by decreasing the fuel supplied to the engine or by applying brakes) and change direction (e.g., by turning the front two wheels).” (Zhu, [0027])
Rationale:
Zhu expressly discloses that computer 110 itself actuates the engine, brakes, and steering in autonomous mode, i.e., automatically controls the vehicle without driver input.
to at least one of reduce a speed of the vehicle,
See at least:
“if the computer 110 determines that the other car is turning at the next intersection as noted above, the computer may slow the vehicle down as it approaches the intersection.” (Zhu, [0051])
Rationale:
Zhu expressly discloses automatically reducing the speed of the vehicle in response to the characterized behavior of another road actor. The limitation is recited in the alternative (“at least one of ... or”), so this alternative alone satisfies the control limitation; the remaining alternatives are mapped below for completeness.
change a lane of the vehicle,
See at least:
“Alternatively, the new control strategy could cause the vehicle 101 to change lanes or to stay a particular distance from the tractor-trailer.” (Zhu, [0059])
Rationale:
Zhu expressly discloses automatically changing a lane of the vehicle in response to the characterized behavior of another road actor.
pull the vehicle onto a road shoulder or a parking area,
See at least:
“If the driver remains unresponsive, the computer may cause the vehicle slow, stop or pull over to a safe location” (Zhu, [0070])
Rationale:
Zhu expressly discloses automatically pulling the vehicle over to a safe location. On a roadway, the safe location to pull over to is a road shoulder or a parking area; those are the only places off the travel lanes where a vehicle can stop. Zhu uses this maneuver for a hazard inside the vehicle (an unresponsive driver), but a PHOSITA would have found it obvious to use the same pull-over maneuver as one of the responses available under Zhu [0109] for keeping passengers away from a dangerously driven vehicle, because removing the vehicle from the travel lanes is a known, finite way to separate it from such a hazard, and it uses a capability the vehicle already has.
increase the speed of the vehicle,
See at least:
“the new control strategy could change the speed of vehicle 101, so as to avoid being directly to the right of tractor-trailer 510.” (Zhu, [0059])
Rationale:
Zhu expressly discloses changing the speed of the vehicle to avoid being alongside the road actor. A speed change has only two directions, and Zhu's computer may cause the vehicle to “accelerate” ([0027]). Increasing the speed is therefore one of two finite, predictable options for getting out from beside the road actor, and would have been obvious to try, for example, when a vehicle is approaching from behind.
overtake the first road actor,
See at least:
“if the computer determines that an object is obstructing the intended path of the vehicle, it may maneuver the vehicle around the obstruction.” (Zhu, [0043])
Rationale:
Zhu expressly discloses automatically maneuvering the vehicle around a road object in its path. When the object is a slower road actor ahead, maneuvering around it and continuing past it is overtaking. Zhu also teaches the component maneuvers, a lane change and a speed change relative to the road actor ([0059]), so overtaking the first road actor is at least an obvious combination of Zhu's disclosed maneuvers applied to the road actor Zhu seeks to avoid.
or increase a distance between the vehicle and the first road actor.
See at least:
“Autonomous vehicles may also keep passengers away from other humans or vehicles (autonomous or not) behaving dangerously” (Zhu, [0109])
Rationale:
Zhu expressly discloses automatically keeping the vehicle and its passengers away from the dangerously driven vehicle, i.e., increasing a distance between the vehicle and the first road actor. Zhu likewise teaches “keeping a greater distance from a certain classification of vehicles” ([0058]).
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
relative to other road actors of the plurality of road actors
a driving style of the first road actor,
Disclosure by Kwon
Kwon discloses:
relative to other road actors of the plurality of road actors
See at least:
“the state information of the vehicle may include a vehicle speed, a vehicle's located lane, a relative distance between the vehicle and a neighboring vehicle, and a relative speed between the vehicle and the neighboring vehicle.” (Kwon, [0044])
Rationale:
Kwon expressly discloses determining, for an observed (target) vehicle, its relative distance and relative speed with respect to a neighboring vehicle, i.e., relative kinematic parameters relative to other road actors. Kwon confirms that this state information belongs to the observed target vehicle (“the driving data may include state information and the action information of the target driving vehicle,” [0065]) and that it is collected by “a sensing module equipped in the driving vehicle” such as “a camera or a LIDAR sensor” ([0042]-[0043]). Applied in Zhu's system, which tracks a plurality of road actors with on-board sensors, the neighboring vehicles are other road actors of that tracked plurality.
a driving style of the first road actor,
See at least:
“The driving characteristic inferring unit 130 infers a driving characteristic coefficient with driving data of the target driving vehicle as an input of the inference model. That is, the driving characteristic inferring unit 130 infers the driving characteristic coefficient using the inference model trained by the model training unit 120 to infer the driving characteristic of the target driving vehicle which is an individual object.” (Kwon, [0064])
Rationale:
Kwon expressly discloses determining a driving characteristic of an individual observed vehicle from that vehicle's driving data. Kwon describes this characteristic as whether a vehicle is “aggressively or stably driven depending on a style of a driver” ([0003]), so the inferred driving characteristic is a driving style. The input driving data is the speed, lane, and relative distance and speed information of [0044], so the style is determined using the kinematic and relative kinematic parameters of the first road actor.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to modify the autonomous driving system of Zhu to (i) determine relative kinematic parameters of a tracked road actor with respect to neighboring road actors and (ii) infer a driving style of that individual road actor from its kinematic and relative kinematic parameters, as taught by Kwon ([0044], [0064]), and then to use that driving style as the basis for Zhu's automatic control response. Zhu predicts a detected vehicle's behavior from behavior information stored “for each classification” of object (Zhu [0053]), so that, for example, every tractor-trailer in a given lane is expected to behave alike (Zhu [0059]). Kwon teaches the advantage of instead inferring “the driving characteristic of the target driving vehicle which is an individual object” (Kwon [0064]), and that the inferred characteristic “may be used to control the learning driving vehicle” (Kwon [0071]). A PHOSITA would have recognized that characterizing how a particular nearby vehicle is actually being driven (its speed, its gaps and closing speeds relative to surrounding vehicles, and its lane position) identifies the “swerving or speeding” vehicles from which Zhu seeks to keep its passengers away (Zhu [0109]) more accurately than a class-level model does, and so improves the safety and accuracy of Zhu's control response. The references are technically compatible: both operate on on-board camera/LIDAR data (Zhu [0032]; Kwon [0043]) in the same field of controlling a vehicle in response to surrounding vehicles. Kwon's inferred style is an input that Zhu's control-strategy selection (Zhu [0059]) can use without changing its principle of operation. The combination is thus the use of a known technique to improve a similar system in the same way, with a reasonable expectation of success. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 417 (2007).
Regarding Claim 2,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 2.
Disclosure by Zhu
Zhu discloses:
wherein the kinematic parameters of the plurality of road actors include at least one of:
See at least:
“The object's current movement may be also be determined by the sensor (e.g., the component is a self-contained speed radar detector) or by the computer 110 based on information provided by the sensors (e.g., by comparing changes in the object's position data over time).” (Zhu, [0042])
Rationale:
Zhu expressly discloses that the kinematic parameters of the tracked road actors comprise their measured movement. The specific alternatives are mapped below. The list is recited in the alternative (“at least one of”), so any one alternative satisfies the claim.
a velocity for each road actor of the plurality of road actors;
See at least:
“the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly discloses the velocity of the tracked object. Zhu tracks each detected object in the same manner ([0042]), so a velocity is determined for each road actor of the plurality.
a yaw turning rate for each road actor of the plurality of road actors;
See at least:
“by comparing changes in the object's position data over time” (Zhu, [0042])
Rationale:
Zhu does not use the term “yaw rate,” but this is obvious from Zhu. Zhu determines each object's “instant direction” ([0044]) by comparing its position data over time. The rate of change of an object's heading over time is, by definition, its yaw (turning) rate. A PHOSITA computing heading from successive positions would have found it obvious to compute its rate of change as well, because the turning rate is the standard kinematic quantity for predicting a vehicle's curved path, which is the same “future movement” prediction Zhu performs ([0044]).
an acceleration of each road actor of the plurality of road actors;
See at least:
“based solely on the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly discloses the acceleration/deceleration of each tracked object.
or a relative position of each road actor with respect to the vehicle.
See at least:
“the lasers may measure the distance between the vehicle and the object surfaces facing the vehicle by spinning on its axis and changing its pitch.” (Zhu, [0035])
Rationale:
Zhu expressly discloses measuring the range from the vehicle to each detected object. That range, together with the object's “location” information ([0042]), is the relative position of each road actor with respect to the vehicle.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 1, with the velocity, direction (yaw), acceleration, and range parameters that Zhu measures for each tracked road actor (Zhu [0035], [0044]) serving as the kinematic parameters supplied to Kwon's driving-style inference. Kwon's inference takes the observed vehicle's speed and related state information as its inputs (Kwon [0044]), so the parameters Zhu already measures for every tracked object are exactly the data Kwon's inference needs. Supplying them yields the predictable result of an individualized driving style for each road actor Zhu tracks, improving the accuracy of Zhu's safety response for the reasons set forth with respect to Claim 1.
Regarding Claim 3,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 3.
Disclosure by Zhu
Zhu discloses:
wherein the one or more relative kinematic parameters of the first road actor include at least one of:
See at least:
“any type of environmental or contextual information may be used to determine the object's current state” (Zhu, [0054])
Rationale:
Zhu expressly discloses that the relative kinematic parameters (the contextual state) of the first road actor may comprise any contextual motion information. The specific alternatives are mapped below; because they are recited in the alternative (“at least one of”), any one satisfies the claim.
or a relative position of the first road actor compared to one or more lane lines.
See at least:
“comparing the GPS position with previously-stored lane-specific map data to determine whether the other car is within a turn lane” (Zhu, [0052])
Rationale:
Zhu expressly discloses determining the position of another vehicle relative to the lanes defined in its map data. That map data identifies “lane lines” ([0052]), so this is a relative position of the first road actor compared to one or more lane lines.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
a relative velocity of the first road actor compared to a velocity of a second road actor of the plurality of road actors;
a relative acceleration of the first road actor compared to an acceleration of the second road actor;
a relative position of the first road actor with respect to the second road actor;
Disclosure by Kwon
Kwon discloses:
a relative velocity of the first road actor compared to a velocity of a second road actor of the plurality of road actors;
See at least:
“a relative speed between the vehicle and the neighboring vehicle” (Kwon, [0044])
Rationale:
Kwon expressly discloses the relative speed of the observed vehicle compared to a neighboring vehicle. Applied in Zhu's tracking of a plurality of road actors, the neighboring vehicle is a second road actor of the plurality.
a relative acceleration of the first road actor compared to an acceleration of the second road actor;
See at least:
“The action information is information regarding a running operation of the vehicle, and may include information related to the lane change and acceleration adjustment.” (Kwon, [0044])
Rationale:
Kwon expressly discloses tracking the observed vehicle's acceleration adjustments, together with its relative speed to a neighboring vehicle, repeatedly over time ([0067]). Kwon does not use the term “relative acceleration,” but the limitation would have been obvious to a PHOSITA on either of two independent grounds. First, Zhu already determines the “acceleration/deceleration” of every object it tracks (Zhu [0044]), including both the first road actor and the second road actor. Comparing two accelerations that are already measured, by taking their difference, is the definition of relative acceleration, and Kwon teaches making exactly this kind of comparison between an observed vehicle and its neighbor for speed and distance ([0044]). Second, the rate of change of the relative speed that Kwon tracks over time is, by definition, the relative acceleration. Either way, the relative acceleration is computed from data the combination already collects, using elementary kinematics. A PHOSITA would have been motivated to compute it because it directly reveals behaviors such as abrupt closing on a lead vehicle or hard braking in front of a follower, which bear on whether a vehicle is being driven dangerously (Zhu [0109]).
a relative position of the first road actor with respect to the second road actor;
See at least:
“a relative distance between the vehicle and a neighboring vehicle” (Kwon, [0044])
Rationale:
Kwon expressly discloses the relative distance, i.e., the relative position, of the observed vehicle with respect to a neighboring (second) road actor.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 1, using as the relative kinematic parameters of the first road actor both Zhu's lane-relative position (Zhu [0052]) and Kwon's relative speed, relative distance, and relative acceleration with respect to a neighboring road actor (Kwon [0044]), the latter obtained by differencing the accelerations Zhu measures for each tracked object (Zhu [0044]) or by differentiating Kwon's tracked relative speed. Kwon identifies these relative quantities as the state information from which an individual vehicle's driving characteristic is inferred (Kwon [0044]). A PHOSITA would have recognized that gaps and closing rates to other vehicles directly reveal tailgating, cutting in, and abrupt closing, which are exactly the dangerous behaviors Zhu seeks to react to (Zhu [0109]). Using them improves the accuracy of the style determination with predictable results.
Regarding Claim 4,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 4.
Disclosure by Zhu
Zhu discloses:
wherein the processing device is configured to:
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses that the processing device is configured, by the stored instructions it executes, to perform the recited functions.
repeat determining the kinematic parameters and relative kinematic parameters of the first road actor
See at least:
“the sensors may continuously update their output to reflect the environment being sensed at or over a range of time, and continuously or as demanded provide that updated output to the computer” (Zhu, [0034])
Rationale:
Zhu expressly discloses continuously updating the sensed data for the tracked objects. Each update necessarily repeats the determination of the object's kinematic parameters and its relative (lane/roadway) parameters, which are determined from that data ([0044], [0054]). In the combination, the relative parameters with respect to other road actors (Kwon) are computed from the same continuously updated data and are repeated with it.
over a time period;
See at least:
“The computer 110 may use the sensor data to continue to track the object's proximity to vehicle 101 (Block 670). If the object remains within a predefined distance of vehicle 101, such as between 0 to 200 meters, then computer 110 may maintain the new control strategy (Block 680).” (Zhu, [0058])
Rationale:
Zhu expressly discloses that the repeated tracking continues over the time period during which the object remains within the predefined distance of the vehicle.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
and determine the driving style of the first actor
using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period.
Disclosure by Kwon
Kwon discloses:
and determine the driving style of the first actor
See at least:
“the driving characteristic inferring unit 130 infers the current driving characteristic coefficient based on additional driving data collected over time as in S20, and repeatedly updates a previous driving characteristic coefficient with the current driving characteristic coefficient.” (Kwon, [0067])
Rationale:
Kwon expressly discloses determining the driving style (driving characteristic) of the observed vehicle from driving data collected over a period of time.
using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period.
See at least:
“the driving characteristic inferring apparatus 100 of the present disclosure may perform the analysis by reflecting information of time-series data which varies over time, as well as a single datum at a specific time to perform more accurate inference.” (Kwon, [0070])
Rationale:
Kwon expressly discloses that the style determination uses the time series of the target vehicle's driving data. That data comprises its speed and its relative distance and relative speed to neighboring vehicles ([0044]), i.e., the kinematic and relative kinematic parameters determined over the time period.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 1, and further to repeat the determination of the first road actor's kinematic and relative kinematic parameters over the period during which Zhu tracks the road actor (Zhu [0058]) and to determine the driving style from the parameters gathered over that period, as taught by Kwon ([0067], [0070]). Kwon teaches that analysis “reflecting information of time-series data which varies over time” performs “more accurate inference” than a single datum (Kwon [0070]), and reports an accuracy improvement of approximately 21% when 10 sequences are used (Kwon [0074]). Zhu already continuously updates its sensor data while the road actor remains within 0 to 200 meters (Zhu [0034], [0058]), so the time series is already available. Using it improves accuracy and reliability with a reasonable expectation of success.
Regarding Claim 8,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 8.
Disclosure by Zhu
Zhu discloses:
wherein the processing device is configured to:
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses that the processing device is configured, by the stored instructions it executes, to perform the recited functions.
generate, using the kinematic parameters and the relative kinematic parameters of the first road actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640).” (Zhu, [0058])
Rationale:
Zhu expressly discloses generating, from the object's speed (a kinematic parameter) and lane position (a relative kinematic parameter), a record of the object's state. In the combination, the relative kinematic parameters with respect to other road actors (Kwon [0044]) are included in that generated record.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
and determine the driving style of the first road actor by providing the driving profile of the first road actor as input to a trained machine learning model,
the trained machine learning model providing the driving style of the first road actor as output responsive to receiving the driving profile of the first road actor as input.
Disclosure by Kwon
Kwon discloses:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the driving data may include state information and the action information of the target driving vehicle, and may be observation result data of the target driving vehicle collected in the monitoring unit 110” (Kwon, [0065])
Rationale:
Kwon expressly discloses generating a collection of driving data for the individual target vehicle. It comprises state information (speed, lane, relative distance and relative speed to neighboring vehicles, [0044]) and action information (lane changes and acceleration adjustments, [0044]), each of which indicates how that vehicle is being driven. The collection is a driving profile of the first road actor including a plurality of metrics indicative of its driving behaviors.
and determine the driving style of the first road actor by providing the driving profile of the first road actor as input to a trained machine learning model,
See at least:
“The real-time inference model training unit 123 trains the driving characteristic inference model in real-time using the generated dataset for every driving characteristic. The inference model may have a recurrent neural network structure ... and specifically, the inference model may include an LSTM network and a dense layer.” (Kwon, [0056])
Rationale:
Kwon expressly discloses a trained machine learning model (an LSTM neural network trained on a labeled dataset). The target vehicle's driving data is provided to that model as its input ([0064]) to determine the driving style.
the trained machine learning model providing the driving style of the first road actor as output responsive to receiving the driving profile of the first road actor as input.
See at least:
“infers the driving characteristic coefficient through the LSTM network and the dense layer.” (Kwon, [0057])
Rationale:
Kwon expressly discloses that the trained model outputs the driving characteristic (driving style) in response to receiving the target vehicle's driving data as input.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 1, and further to generate a driving profile of the first road actor and determine its driving style by providing that profile as input to Kwon's trained LSTM inference model (Kwon [0056]-[0057], [0065]). Kwon teaches that its trained inference model improves inference accuracy over prior inference models (Kwon [0073]). A PHOSITA would have recognized that a trained model can learn complex, time-varying patterns of dangerous driving that are impractical to hand-code, and Zhu itself contemplates machine-learned behavior rules (Zhu [0051], “tools and techniques of machine learning”). Using Kwon's trained model in Zhu's system is a predictable use of a known element according to its established function, and it improves accuracy.
Claim Limitations Not Explicitly Disclosed by the Combination of All References
After combining the teachings of Zhu and Kwon, all limitations of Claim 8 are disclosed or rendered obvious.
Regarding Claim 11,
Disclosure by Zhu
Zhu teaches:
A method for monitoring road actors, the method comprising:
See at least:
“a vehicle implementing the autonomous driving system is capable of detecting and reacting to surrounding objects. Some of the detected objects will be mobile, such as pedestrians, automobiles, and bicycles.” (Zhu, [0017])
Rationale:
Zhu expressly teaches an autonomous driving system whose function is to detect, identify, and react to the mobile objects (automobiles, bicycles, pedestrians) around the vehicle. Those mobile objects are road actors, and continuously detecting and reacting to them is monitoring them. The open-ended transitional term “comprising” permits Zhu's additional components (e.g., navigation and display systems) without affecting the mapping.
receiving, from one or more sensors of a vehicle, sensor data
See at least:
“Upon detecting the object, the vehicle's computer 110 may classify the object based on the data received by the vehicle's sensors (Block 630).” (Zhu, [0058])
Rationale:
Zhu expressly teaches receiving, at computer 110, data from the vehicle's sensors (lasers, radar, cameras, [0032]) concerning detected objects. This is receiving, from one or more sensors of a vehicle, sensor data.
indicative of kinematic parameters of a plurality of road actors
See at least:
“the computer system 110 may simply predict the object's future movement based solely on the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly teaches that the received sensor data yields the direction, acceleration/deceleration, and velocity of each detected object. These are kinematic parameters. Zhu tracks a plurality of such objects at once, “pedestrians, bicycles, other vehicles, or objects in the roadway” ([0042]), so the data is indicative of kinematic parameters of a plurality of road actors.
in a surrounding of the vehicle,
See at least:
“the sensors may provide the location and shape information of objects surrounding the vehicle to computer 110, which in turn may identify the object as another vehicle.” (Zhu, [0042])
Rationale:
Zhu expressly teaches that the road actors whose data is received are “objects surrounding the vehicle,” i.e., in a surrounding of the vehicle.
wherein the plurality of road actors are different from the vehicle;
See at least:
“FIGS. 5A and 5B depict vehicle 101 as driving along side a tractor-trailer 510. As provided in FIG. 5B, the tractor-trailer 510 is within range of several sensor fields of vehicle 101.” (Zhu, [0055])
Rationale:
Zhu expressly teaches that the tracked road actor (tractor-trailer 510) is a separate vehicle driving alongside the host vehicle 101 and detected within the host vehicle's sensor fields. Zhu's tracked objects are, throughout, “objects external to the vehicle” ([0032]); none of them is vehicle 101 itself. The road actors are therefore different from the vehicle.
determining, using the sensor data, relative kinematic parameters of a first road actor of the plurality of road actors
See at least:
“any type of environmental or contextual information may be used to determine the object's current state, such as the object's speed, route the object has traveled, nature of the roadway on which the object is traveling, any lane changes made by the object, or the object's use of headlights or blinkers.” (Zhu, [0054])
Rationale:
Zhu expressly teaches determining, from the sensor data, a state of a particular tracked object (a first road actor of the plurality) that includes its motion in relation to its context: the route it has traveled, its lane changes, and the roadway it travels on. Motion expressed in relation to something other than the host vehicle is a relative kinematic parameter. What that motion is relative to is mapped separately below.
and relative to one or more road features;
See at least:
“data collected from the sensors of vehicle 101 may be used to determine that the tractor-trailer is traveling at 55 miles per hour in the left-hand lane of the interstate.” (Zhu, [0055])
Rationale:
Zhu expressly teaches determining the first road actor's motion relative to road features: its speed within a specific lane (the left-hand lane) of a specific roadway (the interstate). Zhu's stored map includes “lane lines” and “speed limits” for each roadway segment ([0052]), so the object's lane position and speed are determined relative to one or more road features.
determining, using the kinematic parameters and the relative kinematic parameters of the first actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640). Upon determining the objects classification and state, computer 110 accesses behavior model information contained in database 137 (Block 650).” (Zhu, [0058])
Rationale:
Zhu expressly teaches using the object's speed (a kinematic parameter) and its lane position (a relative kinematic parameter with respect to a road feature) as the inputs from which the object's likely behavior is determined. Zhu does not call the output a “driving style”; that limitation is addressed separately below. In the combination, the relative kinematic parameters with respect to other road actors (taught by Kwon, below) are added to these same inputs.
wherein the driving style is indicative of dangerous or reckless driving; and
See at least:
“Autonomous vehicles may also keep passengers away from other humans or vehicles (autonomous or not) behaving dangerously, for example, a swerving or speeding vehicle or distracted a driver talking on a mobile phone.” (Zhu, [0109])
Rationale:
Zhu expressly teaches that the autonomous vehicle characterizes another vehicle's driving as dangerous, with examples (swerving, speeding, distracted driving) that are classic indicators of dangerous or reckless driving. Swerving is evident from the vehicle's lateral motion relative to its lane, and speeding from its velocity relative to the posted limit that Zhu stores for each roadway segment ([0052]). Once Zhu's system determines a driving style of the road actor (as taught by Kwon, below), a PHOSITA would have found it obvious for that style to be one indicative of dangerous or reckless driving, because identifying dangerous drivers is the express purpose for which Zhu characterizes surrounding vehicles in this passage.
based on the driving style of the first road actor,
See at least:
“The new control strategy is therefore specific to the predicted behavior of tractor-trailer 510, in that a different behavior model and a different control strategy would be used if object 510 was a motorcycle or sports car.” (Zhu, [0059])
Rationale:
Zhu expressly teaches selecting the vehicle's control response based on the characterization of the particular road actor's behavior, so that a different characterization produces a different control strategy. Combined with Zhu [0109] (the vehicle reacts to vehicles behaving dangerously) and Kwon's driving style (below), the control is performed based on the driving style of the first road actor.
automatically controlling the vehicle
See at least:
“if the vehicle is operating in a completely autonomous mode, computer 110 may cause the vehicle to accelerate (e.g., by increasing fuel or other energy provided to the engine), decelerate (e.g., by decreasing the fuel supplied to the engine or by applying brakes) and change direction (e.g., by turning the front two wheels).” (Zhu, [0027])
Rationale:
Zhu expressly teaches that computer 110 itself actuates the engine, brakes, and steering in autonomous mode, i.e., automatically controls the vehicle without driver input.
to at least one of reduce a speed of the vehicle,
See at least:
“if the computer 110 determines that the other car is turning at the next intersection as noted above, the computer may slow the vehicle down as it approaches the intersection.” (Zhu, [0051])
Rationale:
Zhu expressly teaches automatically reducing the speed of the vehicle in response to the characterized behavior of another road actor. The limitation is recited in the alternative (“at least one of ... or”), so this alternative alone satisfies the control limitation; the remaining alternatives are mapped below for completeness.
change a lane of the vehicle,
See at least:
“Alternatively, the new control strategy could cause the vehicle 101 to change lanes or to stay a particular distance from the tractor-trailer.” (Zhu, [0059])
Rationale:
Zhu expressly teaches automatically changing a lane of the vehicle in response to the characterized behavior of another road actor.
pull the vehicle onto a road shoulder or a parking area,
See at least:
“If the driver remains unresponsive, the computer may cause the vehicle slow, stop or pull over to a safe location” (Zhu, [0070])
Rationale:
Zhu expressly teaches automatically pulling the vehicle over to a safe location. On a roadway, the safe location to pull over to is a road shoulder or a parking area; those are the only places off the travel lanes where a vehicle can stop. Zhu uses this maneuver for a hazard inside the vehicle (an unresponsive driver), but a PHOSITA would have found it obvious to use the same pull-over maneuver as one of the responses available under Zhu [0109] for keeping passengers away from a dangerously driven vehicle, because removing the vehicle from the travel lanes is a known, finite way to separate it from such a hazard, and it uses a capability the vehicle already has.
increase the speed of the vehicle,
See at least:
“the new control strategy could change the speed of vehicle 101, so as to avoid being directly to the right of tractor-trailer 510.” (Zhu, [0059])
Rationale:
Zhu expressly teaches changing the speed of the vehicle to avoid being alongside the road actor. A speed change has only two directions, and Zhu's computer may cause the vehicle to “accelerate” ([0027]). Increasing the speed is therefore one of two finite, predictable options for getting out from beside the road actor, and would have been obvious to try, for example, when a vehicle is approaching from behind.
overtake the first road actor,
See at least:
“if the computer determines that an object is obstructing the intended path of the vehicle, it may maneuver the vehicle around the obstruction.” (Zhu, [0043])
Rationale:
Zhu expressly teaches automatically maneuvering the vehicle around a road object in its path. When the object is a slower road actor ahead, maneuvering around it and continuing past it is overtaking. Zhu also teaches the component maneuvers, a lane change and a speed change relative to the road actor ([0059]), so overtaking the first road actor is at least an obvious combination of Zhu's disclosed maneuvers applied to the road actor Zhu seeks to avoid.
or increase a distance between the vehicle and the first road actor.
See at least:
“Autonomous vehicles may also keep passengers away from other humans or vehicles (autonomous or not) behaving dangerously” (Zhu, [0109])
Rationale:
Zhu expressly teaches automatically keeping the vehicle and its passengers away from the dangerously driven vehicle, i.e., increasing a distance between the vehicle and the first road actor. Zhu likewise teaches “keeping a greater distance from a certain classification of vehicles” ([0058]).
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
relative to other road actors of the plurality of road actors
a driving style of the first road actor,
Disclosure by Kwon
Kwon teaches:
relative to other road actors of the plurality of road actors
See at least:
“the state information of the vehicle may include a vehicle speed, a vehicle's located lane, a relative distance between the vehicle and a neighboring vehicle, and a relative speed between the vehicle and the neighboring vehicle.” (Kwon, [0044])
Rationale:
Kwon expressly teaches determining, for an observed (target) vehicle, its relative distance and relative speed with respect to a neighboring vehicle, i.e., relative kinematic parameters relative to other road actors. Kwon confirms that this state information belongs to the observed target vehicle (“the driving data may include state information and the action information of the target driving vehicle,” [0065]) and that it is collected by “a sensing module equipped in the driving vehicle” such as “a camera or a LIDAR sensor” ([0042]-[0043]). Applied in Zhu's system, which tracks a plurality of road actors with on-board sensors, the neighboring vehicles are other road actors of that tracked plurality.
a driving style of the first road actor,
See at least:
“The driving characteristic inferring unit 130 infers a driving characteristic coefficient with driving data of the target driving vehicle as an input of the inference model. That is, the driving characteristic inferring unit 130 infers the driving characteristic coefficient using the inference model trained by the model training unit 120 to infer the driving characteristic of the target driving vehicle which is an individual object.” (Kwon, [0064])
Rationale:
Kwon expressly teaches determining a driving characteristic of an individual observed vehicle from that vehicle's driving data. Kwon describes this characteristic as whether a vehicle is “aggressively or stably driven depending on a style of a driver” ([0003]), so the inferred driving characteristic is a driving style. The input driving data is the speed, lane, and relative distance and speed information of [0044], so the style is determined using the kinematic and relative kinematic parameters of the first road actor.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to modify the autonomous driving system of Zhu to (i) determine relative kinematic parameters of a tracked road actor with respect to neighboring road actors and (ii) infer a driving style of that individual road actor from its kinematic and relative kinematic parameters, as taught by Kwon ([0044], [0064]), and then to use that driving style as the basis for Zhu's automatic control response. Zhu predicts a detected vehicle's behavior from behavior information stored “for each classification” of object (Zhu [0053]), so that, for example, every tractor-trailer in a given lane is expected to behave alike (Zhu [0059]). Kwon teaches the advantage of instead inferring “the driving characteristic of the target driving vehicle which is an individual object” (Kwon [0064]), and that the inferred characteristic “may be used to control the learning driving vehicle” (Kwon [0071]). A PHOSITA would have recognized that characterizing how a particular nearby vehicle is actually being driven (its speed, its gaps and closing speeds relative to surrounding vehicles, and its lane position) identifies the “swerving or speeding” vehicles from which Zhu seeks to keep its passengers away (Zhu [0109]) more accurately than a class-level model does, and so improves the safety and accuracy of Zhu's control response. The references are technically compatible: both operate on on-board camera/LIDAR data (Zhu [0032]; Kwon [0043]) in the same field of controlling a vehicle in response to surrounding vehicles. Kwon's inferred style is an input that Zhu's control-strategy selection (Zhu [0059]) can use without changing its principle of operation. The combination is thus the use of a known technique to improve a similar system in the same way, with a reasonable expectation of success. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 417 (2007).
Regarding Claim 12,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 12.
Disclosure by Zhu
Zhu teaches:
wherein the kinematic parameters of the plurality of road actors include at least one of:
See at least:
“The object's current movement may be also be determined by the sensor (e.g., the component is a self-contained speed radar detector) or by the computer 110 based on information provided by the sensors (e.g., by comparing changes in the object's position data over time).” (Zhu, [0042])
Rationale:
Zhu expressly teaches that the kinematic parameters of the tracked road actors comprise their measured movement. The specific alternatives are mapped below. The list is recited in the alternative (“at least one of”), so any one alternative satisfies the claim.
a velocity for each road actor of the plurality of road actors;
See at least:
“the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly teaches the velocity of the tracked object. Zhu tracks each detected object in the same manner ([0042]), so a velocity is determined for each road actor of the plurality.
a yaw tuning rate for each road actor of the plurality of road actors;
See at least:
“by comparing changes in the object's position data over time” (Zhu, [0042])
Rationale:
Zhu does not use the term “yaw rate,” but this is obvious from Zhu. Zhu determines each object's “instant direction” ([0044]) by comparing its position data over time. The rate of change of an object's heading over time is, by definition, its yaw (turning) rate. A PHOSITA computing heading from successive positions would have found it obvious to compute its rate of change as well, because the turning rate is the standard kinematic quantity for predicting a vehicle's curved path, which is the same “future movement” prediction Zhu performs ([0044]).
an acceleration of each road actor of the plurality of road actors;
See at least:
“based solely on the object's instant direction, acceleration/deceleration and velocity” (Zhu, [0044])
Rationale:
Zhu expressly teaches the acceleration/deceleration of each tracked object.
or a relative position of each road actor with respect to the vehicle.
See at least:
“the lasers may measure the distance between the vehicle and the object surfaces facing the vehicle by spinning on its axis and changing its pitch.” (Zhu, [0035])
Rationale:
Zhu expressly teaches measuring the range from the vehicle to each detected object. That range, together with the object's “location” information ([0042]), is the relative position of each road actor with respect to the vehicle.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, with the velocity, direction (yaw), acceleration, and range parameters that Zhu measures for each tracked road actor (Zhu [0035], [0044]) serving as the kinematic parameters supplied to Kwon's driving-style inference. Kwon's inference takes the observed vehicle's speed and related state information as its inputs (Kwon [0044]), so the parameters Zhu already measures for every tracked object are exactly the data Kwon's inference needs. Supplying them yields the predictable result of an individualized driving style for each road actor Zhu tracks, improving the accuracy of Zhu's safety response for the reasons set forth with respect to Claim 11.
Regarding Claim 13,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 13.
Disclosure by Zhu
Zhu teaches:
wherein the one or more relative kinematic parameters of the first road actor include at least one of:
See at least:
“any type of environmental or contextual information may be used to determine the object's current state” (Zhu, [0054])
Rationale:
Zhu expressly teaches that the relative kinematic parameters (the contextual state) of the first road actor may comprise any contextual motion information. The specific alternatives are mapped below; because they are recited in the alternative (“at least one of”), any one satisfies the claim.
or a relative position of the first road actor compared to one or more lane lines.
See at least:
“comparing the GPS position with previously-stored lane-specific map data to determine whether the other car is within a turn lane” (Zhu, [0052])
Rationale:
Zhu expressly teaches determining the position of another vehicle relative to the lanes defined in its map data. That map data identifies “lane lines” ([0052]), so this is a relative position of the first road actor compared to one or more lane lines.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
a relative velocity of the first road actor compared to a velocity of a second road actor of the plurality of road actors;
a relative acceleration of the first road actor compared to an acceleration of the second road actor;
a relative position of the first road actor with respect to the second road actor;
Disclosure by Kwon
Kwon teaches:
a relative velocity of the first road actor compared to a velocity of a second road actor of the plurality of road actors;
See at least:
“a relative speed between the vehicle and the neighboring vehicle” (Kwon, [0044])
Rationale:
Kwon expressly teaches the relative speed of the observed vehicle compared to a neighboring vehicle. Applied in Zhu's tracking of a plurality of road actors, the neighboring vehicle is a second road actor of the plurality.
a relative acceleration of the first road actor compared to an acceleration of the second road actor;
See at least:
“The action information is information regarding a running operation of the vehicle, and may include information related to the lane change and acceleration adjustment.” (Kwon, [0044])
Rationale:
Kwon expressly teaches tracking the observed vehicle's acceleration adjustments, together with its relative speed to a neighboring vehicle, repeatedly over time ([0067]). Kwon does not use the term “relative acceleration,” but the limitation would have been obvious to a PHOSITA on either of two independent grounds. First, Zhu already determines the “acceleration/deceleration” of every object it tracks (Zhu [0044]), including both the first road actor and the second road actor. Comparing two accelerations that are already measured, by taking their difference, is the definition of relative acceleration, and Kwon teaches making exactly this kind of comparison between an observed vehicle and its neighbor for speed and distance ([0044]). Second, the rate of change of the relative speed that Kwon tracks over time is, by definition, the relative acceleration. Either way, the relative acceleration is computed from data the combination already collects, using elementary kinematics. A PHOSITA would have been motivated to compute it because it directly reveals behaviors such as abrupt closing on a lead vehicle or hard braking in front of a follower, which bear on whether a vehicle is being driven dangerously (Zhu [0109]).
a relative position of the first road actor with respect to the second road actor;
See at least:
“a relative distance between the vehicle and a neighboring vehicle” (Kwon, [0044])
Rationale:
Kwon expressly teaches the relative distance, i.e., the relative position, of the observed vehicle with respect to a neighboring (second) road actor.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, using as the relative kinematic parameters of the first road actor both Zhu's lane-relative position (Zhu [0052]) and Kwon's relative speed, relative distance, and relative acceleration with respect to a neighboring road actor (Kwon [0044]), the latter obtained by differencing the accelerations Zhu measures for each tracked object (Zhu [0044]) or by differentiating Kwon's tracked relative speed. Kwon identifies these relative quantities as the state information from which an individual vehicle's driving characteristic is inferred (Kwon [0044]). A PHOSITA would have recognized that gaps and closing rates to other vehicles directly reveal tailgating, cutting in, and abrupt closing, which are exactly the dangerous behaviors Zhu seeks to react to (Zhu [0109]). Using them improves the accuracy of the style determination with predictable results.
Regarding Claim 14,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 14.
Disclosure by Zhu
Zhu teaches:
repeating determining the kinematic parameters and relative kinematic parameters of the first road actor
See at least:
“the sensors may continuously update their output to reflect the environment being sensed at or over a range of time, and continuously or as demanded provide that updated output to the computer” (Zhu, [0034])
Rationale:
Zhu expressly teaches continuously updating the sensed data for the tracked objects. Each update necessarily repeats the determination of the object's kinematic parameters and its relative (lane/roadway) parameters, which are determined from that data ([0044], [0054]). In the combination, the relative parameters with respect to other road actors (Kwon) are computed from the same continuously updated data and are repeated with it.
over a time period;
See at least:
“The computer 110 may use the sensor data to continue to track the object's proximity to vehicle 101 (Block 670). If the object remains within a predefined distance of vehicle 101, such as between 0 to 200 meters, then computer 110 may maintain the new control strategy (Block 680).” (Zhu, [0058])
Rationale:
Zhu expressly teaches that the repeated tracking continues over the time period during which the object remains within the predefined distance of the vehicle.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
and determining the driving style of the first road actor
using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period.
Disclosure by Kwon
Kwon teaches:
and determining the driving style of the first road actor
See at least:
“the driving characteristic inferring unit 130 infers the current driving characteristic coefficient based on additional driving data collected over time as in S20, and repeatedly updates a previous driving characteristic coefficient with the current driving characteristic coefficient.” (Kwon, [0067])
Rationale:
Kwon expressly teaches determining the driving style (driving characteristic) of the observed vehicle from driving data collected over a period of time.
using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period.
See at least:
“the driving characteristic inferring apparatus 100 of the present disclosure may perform the analysis by reflecting information of time-series data which varies over time, as well as a single datum at a specific time to perform more accurate inference.” (Kwon, [0070])
Rationale:
Kwon expressly teaches that the style determination uses the time series of the target vehicle's driving data. That data comprises its speed and its relative distance and relative speed to neighboring vehicles ([0044]), i.e., the kinematic and relative kinematic parameters determined over the time period.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, and further to repeat the determination of the first road actor's kinematic and relative kinematic parameters over the period during which Zhu tracks the road actor (Zhu [0058]) and to determine the driving style from the parameters gathered over that period, as taught by Kwon ([0067], [0070]). Kwon teaches that analysis “reflecting information of time-series data which varies over time” performs “more accurate inference” than a single datum (Kwon [0070]), and reports an accuracy improvement of approximately 21% when 10 sequences are used (Kwon [0074]). Zhu already continuously updates its sensor data while the road actor remains within 0 to 200 meters (Zhu [0034], [0058]), so the time series is already available. Using it improves accuracy and reliability with a reasonable expectation of success.
Regarding Claim 18,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 18.
Disclosure by Zhu
Zhu teaches:
generating, using the kinematic parameters and the relative kinematic parameters of the first road actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640).” (Zhu, [0058])
Rationale:
Zhu expressly teaches generating, from the object's speed (a kinematic parameter) and lane position (a relative kinematic parameter), a record of the object's state. In the combination, the relative kinematic parameters with respect to other road actors (Kwon [0044]) are included in that generated record.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
a driving profile of the road first actor including a plurality of metrics indicative of driving behaviors of the first road actor;
and determining the driving style of the first road actor by providing the driving profile of the first road actor as input to a trained machine learning model,
the trained machine learning model providing the driving style of the first road actor as output responsive to receiving the driving profile of the first road actor as input.
Disclosure by Kwon
Kwon teaches:
a driving profile of the road first actor including a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the driving data may include state information and the action information of the target driving vehicle, and may be observation result data of the target driving vehicle collected in the monitoring unit 110” (Kwon, [0065])
Rationale:
Kwon expressly teaches generating a collection of driving data for the individual target vehicle. It comprises state information (speed, lane, relative distance and relative speed to neighboring vehicles, [0044]) and action information (lane changes and acceleration adjustments, [0044]), each of which indicates how that vehicle is being driven. The collection is a driving profile of the first road actor including a plurality of metrics indicative of its driving behaviors.
and determining the driving style of the first road actor by providing the driving profile of the first road actor as input to a trained machine learning model,
See at least:
“The real-time inference model training unit 123 trains the driving characteristic inference model in real-time using the generated dataset for every driving characteristic. The inference model may have a recurrent neural network structure ... and specifically, the inference model may include an LSTM network and a dense layer.” (Kwon, [0056])
Rationale:
Kwon expressly teaches a trained machine learning model (an LSTM neural network trained on a labeled dataset). The target vehicle's driving data is provided to that model as its input ([0064]) to determine the driving style.
the trained machine learning model providing the driving style of the first road actor as output responsive to receiving the driving profile of the first road actor as input.
See at least:
“infers the driving characteristic coefficient through the LSTM network and the dense layer.” (Kwon, [0057])
Rationale:
Kwon expressly teaches that the trained model outputs the driving characteristic (driving style) in response to receiving the target vehicle's driving data as input.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, and further to generate a driving profile of the first road actor and determine its driving style by providing that profile as input to Kwon's trained LSTM inference model (Kwon [0056]-[0057], [0065]). Kwon teaches that its trained inference model improves inference accuracy over prior inference models (Kwon [0073]). A PHOSITA would have recognized that a trained model can learn complex, time-varying patterns of dangerous driving that are impractical to hand-code, and Zhu itself contemplates machine-learned behavior rules (Zhu [0051], “tools and techniques of machine learning”). Using Kwon's trained model in Zhu's system is a predictable use of a known element according to its established function, and it improves accuracy.
Regarding Claim 20,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 20.
Disclosure by Zhu
Zhu teaches:
wherein the one or more actions to be taken by the vehicle
See at least:
“The computer may cause the vehicle to take particular actions in response to the predicted actions of the surrounding objects.” (Zhu, [0051])
Rationale:
As best understood (see the 35 U.S.C. 112(b) rejection of this claim), “the one or more actions to be taken by the vehicle” is interpreted as the control actions recited in Claim 11. Zhu expressly teaches actions taken by the vehicle in response to the characterized behavior of a road actor.
depend on the relative position of the first road actor with respect to the vehicle,
See at least:
“the new control strategy could change the speed of vehicle 101, so as to avoid being directly to the right of tractor-trailer 510.” (Zhu, [0059])
Rationale:
As best understood, “the relative position” is interpreted as a relative position of the first road actor with respect to the vehicle. Zhu expressly teaches selecting the action according to where the road actor is relative to the vehicle (avoiding being directly to its right), i.e., the actions depend on that relative position.
the one or more actions including at least one of”
See at least:
“Vehicle 101 could then implement a new control strategy in response to the accessed behavior model.” (Zhu, [0059])
Rationale:
Zhu expressly teaches a set of available actions, of which at least one is implemented. The specific alternatives are mapped below; because they are recited in the alternative, any one satisfies the claim.
reduce vehicle speed;
See at least:
“if the computer 110 determines that the other car is turning at the next intersection as noted above, the computer may slow the vehicle down as it approaches the intersection.” (Zhu, [0051])
Rationale:
Zhu expressly teaches reducing vehicle speed in response to the road actor's characterized behavior.
change lanes;
See at least:
“Alternatively, the new control strategy could cause the vehicle 101 to change lanes or to stay a particular distance from the tractor-trailer.” (Zhu, [0059])
Rationale:
Zhu expressly teaches changing lanes in response to the road actor's characterized behavior.
or pull into a road shoulder.
See at least:
“If the driver remains unresponsive, the computer may cause the vehicle slow, stop or pull over to a safe location” (Zhu, [0070])
Rationale:
Zhu expressly teaches pulling the vehicle over to a safe location, which on a roadway is the road shoulder. A PHOSITA would have found it obvious to use this existing pull-over maneuver as a response to a dangerous road actor under Zhu [0109], for the reasons given for Claim 11.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, in which Zhu's control actions (reducing speed, changing lanes, pulling over) are selected according to the road actor's position relative to the vehicle (Zhu [0059]) once the road actor's driving style has been determined as taught by Kwon ([0064]). A PHOSITA would have recognized that the safe response to a dangerous driver depends on where that driver is (slowing is safe when the driver is ahead, but not when the driver is closing from behind), and Zhu already selects its response on that basis. The combination therefore yields predictable results and improves the safety of the control response.
Rejection II: Zhu in view of Kwon, and further in view of Huang
Claims 5–7, 9, 15–17, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Zhu in view of Kwon, and further in view of Huang et al. (US 2010/0023223 A1, hereinafter “Huang”).
Regarding Claim 5,
The combination of Zhu and Kwon establishes the system of Claim 4, which is the basis for Claim 5.
Disclosure by Zhu
Zhu discloses:
wherein the processing device is configured to:
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses that the processing device is configured, by the stored instructions it executes, to perform the recited functions.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
determine, using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period,
a plurality of metrics indicative of driving behaviors of the first road actor;
and determine the driving style of the first road actor using the plurality of metrics.
Disclosure by Kwon
Kwon discloses:
determine, using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period,
See at least:
“infers the current driving characteristic coefficient based on additional driving data collected over time” (Kwon, [0067])
Rationale:
Kwon expressly discloses processing the kinematic and relative kinematic parameters of the observed vehicle that were collected over the time period ([0044]). The specific output of that processing, a plurality of metrics, is addressed by Huang below.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 4, in which the kinematic and relative kinematic parameters of the first road actor collected over the time period are processed to determine its driving style, as taught by Kwon ([0067]), for the accuracy and safety reasons set forth with respect to Claims 1 and 4.
Claim Limitations Not Explicitly Disclosed by the Combination of Zhu and Kwon
After combining the teachings of Zhu and Kwon, the following claim limitations are not explicitly disclosed:
a plurality of metrics indicative of driving behaviors of the first road actor;
and determine the driving style of the first road actor using the plurality of metrics.
Disclosure by Huang
Huang discloses:
a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the average headway distance, or headway time, the average velocity of the vehicle, the traffic index ..., and the road condition index ..., including the road type index and ambient condition index, are used as the original features in the classification.” (Huang, [0130])
Rationale:
Huang expressly discloses computing a plurality of metrics (average headway distance, average velocity) from kinematic and relative kinematic data gathered over a driving period. Huang states that “the headway distance/time is a better reflection of the driver's driving style” ([0130]), so these metrics are indicative of driving behaviors. Huang computes them for the host vehicle. In the combination they are computed for the first road actor from the parameters Zhu and Kwon already track for it, as explained in the motivation below.
and determine the driving style of the first road actor using the plurality of metrics.
See at least:
“A neural network can be designed for the classification where the network has an input layer with five input neurons corresponding to the five discriminants, a hidden layer and an output layer with 1 neuron. The output of the net ranges from 1-5, with 1 indicating a rather conservative driver, 3 a typical driver and 5 a rather sporty driver.” (Huang, [0130])
Rationale:
Huang expressly discloses determining the driving style from the plurality of metrics: the metrics are the classifier inputs and the style is the output.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the first road actor's kinematic and relative kinematic parameters collected over the time period are reduced to driving-behavior metrics of the type taught by Huang (for example, average headway distance, average gaps to adjacent-lane vehicles, jerk, and deceleration response, Huang [0076], [0130], [0134], [0158]), and so that the driving style is determined from those metrics. Huang teaches that such metrics are “a better reflection of the driver's driving style” than raw signals (Huang [0130]). Huang also teaches that, because such signals “are time-domain series, data reduction is necessary in order to reduce the complexity of the classifier” (Huang [0131]), and that the resulting style characterization “can be used in various vehicle control algorithms” (Huang [0050]). A PHOSITA would also have recognized that supplying such reduced metrics, rather than raw high-rate signals, to the trained inference model of Kwon ([0056]) gives that model a compact, fixed-length input, which lowers the computation needed for real-time inference on board the vehicle and improves the stability of its output. Kwon already collects the observed vehicle's relative distance and relative speed to neighboring vehicles over time (Kwon [0044], [0067]). Huang's metrics are computed from generic kinematic quantities (range, range rate, speed, acceleration, jerk) that are as available for an observed vehicle as for the host vehicle. Applying them to the road actor that Zhu and Kwon track therefore does not change their principle of operation, and it predictably yields a more stable, lower-complexity style determination. That improves the efficiency, accuracy, and reliability with which a dangerous road actor is identified for Zhu's safety response (Zhu [0109]).
Regarding Claim 6,
The combination of Zhu, Kwon, and Huang establishes the system of Claim 5, which is the basis for Claim 6.
Disclosure by Zhu
Zhu discloses:
a number of lane-line crossings;
See at least:
“any lane changes made by the object” (Zhu, [0054])
Rationale:
Zhu expressly discloses tracking the lane changes made by the observed road actor. Every lane change crosses a lane line. Counting the lane changes recorded over the tracking period gives a number of lane-line crossings, and a PHOSITA would have found it obvious to count them, since frequent lane changes are a recognized sign of the “swerving” behavior that Zhu identifies as dangerous ([0109]).
or violations of driving rules over the time period.
See at least:
“the predicted behavior of other objects is based not only on the type of object and its current trajectory, but also based on some likelihood that the object may obey traffic rules or pre-determined behaviors.” (Zhu, [0051])
Rationale:
Zhu expressly discloses evaluating whether a tracked road actor obeys traffic rules, and identifies “speeding” (a driving rule violation) as dangerous behavior ([0109]). Recording these violations over the tracking period of [0058] gives violations of driving rules over the time period.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
wherein the plurality of metrics include at least one of:
an average distance between the first road actor and a second road actor in front of the first road actor;
an average distance between the first road actor and a lateral road actor;
an average jerk value;
an average delay in deceleration response;
Disclosure by Huang
Huang discloses:
wherein the plurality of metrics include at least one of:
See at least:
“During steady-state vehicle following, the driver's main purpose in headway control is to maintain his or her headway distance or headway time” (Huang, [0130])
Rationale:
Huang expressly discloses that the metrics used for style classification are drawn from distances, speeds, and accelerations of the observed vehicle relative to surrounding vehicles. The specific alternatives are mapped below; because they are recited in the alternative (“at least one of”), any one satisfies the claim. Zhu, as noted above, already discloses two of them.
an average distance between the first road actor and a second road actor in front of the first road actor;
See at least:
“the average headway distance, or headway time, the average velocity of the vehicle ... are used as the original features in the classification.” (Huang, [0130])
Rationale:
Huang expressly discloses the average headway distance, i.e., the average distance between a vehicle and the road actor in front of it. In the combination, this metric is computed for the first road actor and its preceding vehicle (a second road actor), using the relative distance Kwon tracks between the observed vehicle and its neighboring vehicle (Kwon [0044]).
an average distance between the first road actor and a lateral road actor;
See at least:
“The range variable R can be estimated as a weighted average between the headway distance ... and the running average of the adjacent lane vehicle gaps” (Huang, [0076])
Rationale:
Huang expressly discloses computing a running average of the gaps to vehicles in adjacent lanes, i.e., an average distance to lateral road actors. Huang uses this average for its traffic index, which is itself an input to style classification ([0096]). Computing the same running average for the first road actor and a lateral road actor is a direct application of Huang's computation to the tracked vehicle.
an average jerk value;
See at least:
“From the lateral motion point of view, the steering angle, yaw rate, lateral acceleration and lateral jerk can all reflect a driver's driving style.” (Huang, [0158])
Rationale:
Huang expressly discloses that jerk reflects driving style, and uses jerk values as style features (e.g., the maximum lateral jerk, [0163]). Huang uses the maximum rather than the average, but Huang summarizes other signals by their averages (“average speed,” [0165]; “average headway distance,” [0130]). Choosing the average rather than the maximum is a choice among a small number of standard statistics that Huang itself uses interchangeably, so an average jerk value would have been obvious to a PHOSITA; the average is the more robust choice against single-sample noise.
an average delay in deceleration response;
See at least:
“When another vehicle cuts in, the subject vehicle usually decelerates until the headway distance/time reaches the steady-state headway distance/time preferred by the driver. A more conservative driver usually decelerates faster to get back to his/her comfort level quicker, while a sportier driver has a higher tolerance of the shorter distance and decelerates relatively slowly.” (Huang, [0134])
Rationale:
Huang expressly discloses using how quickly a driver decelerates in response to a reduced gap, including “the time to reach the preferred headway time” ([0134]), as a driving-style feature. That is a delay in deceleration response. Averaging it over the cut-in events in a time period is the same averaging Huang applies to other features ([0130]), so an average delay in deceleration response would have been obvious to a PHOSITA.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the first road actor's kinematic and relative kinematic parameters collected over the time period are reduced to driving-behavior metrics of the type taught by Huang (for example, average headway distance, average gaps to adjacent-lane vehicles, jerk, and deceleration response, Huang [0076], [0130], [0134], [0158]), and so that the driving style is determined from those metrics. Huang teaches that such metrics are “a better reflection of the driver's driving style” than raw signals (Huang [0130]). Huang also teaches that, because such signals “are time-domain series, data reduction is necessary in order to reduce the complexity of the classifier” (Huang [0131]), and that the resulting style characterization “can be used in various vehicle control algorithms” (Huang [0050]). A PHOSITA would also have recognized that supplying such reduced metrics, rather than raw high-rate signals, to the trained inference model of Kwon ([0056]) gives that model a compact, fixed-length input, which lowers the computation needed for real-time inference on board the vehicle and improves the stability of its output. Kwon already collects the observed vehicle's relative distance and relative speed to neighboring vehicles over time (Kwon [0044], [0067]). Huang's metrics are computed from generic kinematic quantities (range, range rate, speed, acceleration, jerk) that are as available for an observed vehicle as for the host vehicle. Applying them to the road actor that Zhu and Kwon track therefore does not change their principle of operation, and it predictably yields a more stable, lower-complexity style determination. That improves the efficiency, accuracy, and reliability with which a dangerous road actor is identified for Zhu's safety response (Zhu [0109]).
Regarding Claim 7,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 7.
Disclosure by Zhu
Zhu discloses:
wherein the processing device is configured to:
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses that the processing device is configured, by the stored instructions it executes, to perform the recited functions.
generate, using the kinematic parameters and the relative kinematic parameters of the first road actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640).” (Zhu, [0058])
Rationale:
Zhu expressly discloses generating, from the object's speed (a kinematic parameter) and lane position (a relative kinematic parameter), a record of the object's state. In the combination, the relative kinematic parameters with respect to other road actors (Kwon [0044]) are included in that generated record.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
compare the driving profile of the first road actor to one or more reference driving profiles;
and determine the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
Disclosure by Kwon
Kwon discloses:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the driving data may include state information and the action information of the target driving vehicle, and may be observation result data of the target driving vehicle collected in the monitoring unit 110” (Kwon, [0065])
Rationale:
Kwon expressly discloses generating a collection of driving data for the individual target vehicle. It comprises state information (speed, lane, relative distance and relative speed to neighboring vehicles, [0044]) and action information (lane changes and acceleration adjustments, [0044]), each of which indicates how that vehicle is being driven. The collection is a driving profile of the first road actor including a plurality of metrics indicative of its driving behaviors.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 1, and further to organize the first road actor's kinematic and relative kinematic parameters, together with its action information, into a driving profile of that individual road actor as taught by Kwon ([0065]). This gives the style determination a single structured input per road actor, which is a predictable arrangement of known data that improves the reliability of the per-actor determination for the reasons set forth with respect to Claim 1.
Claim Limitations Not Explicitly Disclosed by the Combination of Zhu and Kwon
After combining the teachings of Zhu and Kwon, the following claim limitations are not explicitly disclosed:
compare the driving profile of the first road actor to one or more reference driving profiles;
and determine the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
Disclosure by Huang
Huang discloses:
compare the driving profile of the first road actor to one or more reference driving profiles;
See at least:
“The basic idea of the FCM-based classification is to determine the class of a pattern, which is represented by a discriminant vector, based on its distance to each pre-determined cluster center. ... C is the total number of pre-determined clusters, such as C=3-5 representing the different levels of sporty driving.” (Huang, [0173])
Rationale:
Huang expressly discloses comparing a driving profile (the “discriminant vector” of driving-behavior metrics) to a set of pre-determined cluster centers. Each cluster center is a reference profile representing a defined level of driving style, so the comparison is to one or more reference driving profiles.
and determine the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
See at least:
“Based on the distances, the algorithm further determines the membership degree” (Huang, [0174])
Rationale:
Huang expressly discloses determining the style class from the computed distances to the reference cluster centers, i.e., determining the driving style based on the comparison of the driving profile to the one or more reference driving profiles.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon to determine the first road actor's driving style by comparing its driving profile with pre-determined reference profiles, as taught by Huang's classification against pre-determined cluster centers that represent defined driving-style levels (Huang [0173]-[0174]). Huang expressly teaches that “fuzzy logic, clustering, neural networks (NN), self-organizing maps (SOM), and even simple threshold-base logic” are alternative techniques for classifying driving style (Huang [0103]). Using Huang's reference-profile comparison in place of, or in addition to, Kwon's inference model is therefore a simple substitution of one known style classifier for another, with predictable results. A PHOSITA would have been motivated to make it for three reasons. First, the comparison yields a membership degree in each style class (Huang [0174]), so the system can directly gauge how closely an observed road actor matches a dangerous profile. That improves the transparency and reliability of Zhu's control decision. Second, the reference profiles are “determined during the design phase” (Huang [0173]), which reduces on-board computation. Third, the predefined style classes can be defined to correspond to the dangerous driving Zhu seeks to avoid (Zhu [0109]).
Regarding Claim 9,
The combination of Zhu and Kwon establishes the system of Claim 1, which is the basis for Claim 9.
Disclosure by Zhu
Zhu discloses:
wherein the processing device is configured to:
See at least:
“The memory 130 stores information accessible by processor 120, including instructions 132 and data 134 that may be executed or otherwise used by the processor 120.” (Zhu, [0019])
Rationale:
Zhu expressly discloses that the processing device is configured, by the stored instructions it executes, to perform the recited functions.
obtain information indicative of at least one of a road condition or weather condition;
See at least:
“if the vehicle determines that another object is a bicycle that is beginning to ascend a steep hill in front of the vehicle, the computer may predict that the bicycle will soon slow down” (Zhu, [0048])
Rationale:
Zhu expressly discloses obtaining information indicative of a road condition (a steep grade on the road ahead) and using it to evaluate a road actor's behavior. The limitation is recited in the alternative (“at least one of a road condition or weather condition”), so the road-condition alternative satisfies it.
Claim Limitations Not Explicitly Disclosed by Zhu
Zhu does not explicitly disclose the following claim limitations:
and determine the driving style of the first actor further based on the information indicative of the at least one of the road condition or weather condition.
Disclosure by Huang
Huang discloses:
and determine the driving style of the first actor further based on the information indicative of the at least one of the road condition or weather condition.
See at least:
“the style classification processor 52 classifies driving style based on driver inputs and vehicle motion, as well as the traffic/road conditions. That is, the traffic/road condition indexes are part of the discriminant features (discussed below) used in the style classification.” (Huang, [0096])
Rationale:
Huang expressly discloses determining the driving style further based on information indicative of road conditions. Huang's road condition index includes weather (“weather condition, such as fog, rain, snow,” [0089]), so the determination is based on the road condition, the weather condition, or both.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the driving-style determination for the first road actor is further based on road and weather information, as taught by Huang ([0089], [0096]). Huang teaches making “the traffic/road condition indexes ... part of the discriminant features” used in style classification (Huang [0096]), including ambient conditions such as “fog, rain, snow” (Huang [0089]), because conditions determine “whether the current driving behavior is valuable for the characterization” (Huang [0097]). A PHOSITA would have recognized that the same behavior can be ordinary in one condition and dangerous in another; for example, a following distance that is acceptable on dry pavement may be reckless in rain. Conditioning the style determination on such information therefore improves its accuracy and reduces both false and missed identifications of dangerous road actors, improving the safety of Zhu's control response. Zhu already obtains road information (Zhu [0048]), and Huang's indexes are simply additional classifier inputs. The combination is thus a predictable use of known elements according to their established functions.
Regarding Claim 15,
The combination of Zhu and Kwon establishes the method of Claim 14, which is the basis for Claim 15.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
determining, using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period,
a plurality of metrics indicative of driving behaviors of the first road actor;
and determining the driving style of the first road actor using the plurality of metrics.
Disclosure by Kwon
Kwon teaches:
determining, using the kinematic parameters and relative kinematic parameters of the first road actor determined over the time period,
See at least:
“infers the current driving characteristic coefficient based on additional driving data collected over time” (Kwon, [0067])
Rationale:
Kwon expressly teaches processing the kinematic and relative kinematic parameters of the observed vehicle that were collected over the time period ([0044]). The specific output of that processing, a plurality of metrics, is addressed by Huang below.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 14, in which the kinematic and relative kinematic parameters of the first road actor collected over the time period are processed to determine its driving style, as taught by Kwon ([0067]), for the accuracy and safety reasons set forth with respect to Claims 11 and 14.
Claim Limitations Not Explicitly Taught by the Combination of Zhu and Kwon
After combining the teachings of Zhu and Kwon, the following claim limitations are not explicitly taught:
a plurality of metrics indicative of driving behaviors of the first road actor;
and determining the driving style of the first road actor using the plurality of metrics.
Disclosure by Huang
Huang teaches:
a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the average headway distance, or headway time, the average velocity of the vehicle, the traffic index ..., and the road condition index ..., including the road type index and ambient condition index, are used as the original features in the classification.” (Huang, [0130])
Rationale:
Huang expressly teaches computing a plurality of metrics (average headway distance, average velocity) from kinematic and relative kinematic data gathered over a driving period. Huang states that “the headway distance/time is a better reflection of the driver's driving style” ([0130]), so these metrics are indicative of driving behaviors. Huang computes them for the host vehicle. In the combination they are computed for the first road actor from the parameters Zhu and Kwon already track for it, as explained in the motivation below.
and determining the driving style of the first road actor using the plurality of metrics.
See at least:
“A neural network can be designed for the classification where the network has an input layer with five input neurons corresponding to the five discriminants, a hidden layer and an output layer with 1 neuron. The output of the net ranges from 1-5, with 1 indicating a rather conservative driver, 3 a typical driver and 5 a rather sporty driver.” (Huang, [0130])
Rationale:
Huang expressly teaches determining the driving style from the plurality of metrics: the metrics are the classifier inputs and the style is the output.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the first road actor's kinematic and relative kinematic parameters collected over the time period are reduced to driving-behavior metrics of the type taught by Huang (for example, average headway distance, average gaps to adjacent-lane vehicles, jerk, and deceleration response, Huang [0076], [0130], [0134], [0158]), and so that the driving style is determined from those metrics. Huang teaches that such metrics are “a better reflection of the driver's driving style” than raw signals (Huang [0130]). Huang also teaches that, because such signals “are time-domain series, data reduction is necessary in order to reduce the complexity of the classifier” (Huang [0131]), and that the resulting style characterization “can be used in various vehicle control algorithms” (Huang [0050]). A PHOSITA would also have recognized that supplying such reduced metrics, rather than raw high-rate signals, to the trained inference model of Kwon ([0056]) gives that model a compact, fixed-length input, which lowers the computation needed for real-time inference on board the vehicle and improves the stability of its output. Kwon already collects the observed vehicle's relative distance and relative speed to neighboring vehicles over time (Kwon [0044], [0067]). Huang's metrics are computed from generic kinematic quantities (range, range rate, speed, acceleration, jerk) that are as available for an observed vehicle as for the host vehicle. Applying them to the road actor that Zhu and Kwon track therefore does not change their principle of operation, and it predictably yields a more stable, lower-complexity style determination. That improves the efficiency, accuracy, and reliability with which a dangerous road actor is identified for Zhu's safety response (Zhu [0109]).
Regarding Claim 16,
The combination of Zhu, Kwon, and Huang establishes the method of Claim 15, which is the basis for Claim 16.
Disclosure by Zhu
Zhu teaches:
a number of lane-line crossings;
See at least:
“any lane changes made by the object” (Zhu, [0054])
Rationale:
Zhu expressly teaches tracking the lane changes made by the observed road actor. Every lane change crosses a lane line. Counting the lane changes recorded over the tracking period gives a number of lane-line crossings, and a PHOSITA would have found it obvious to count them, since frequent lane changes are a recognized sign of the “swerving” behavior that Zhu identifies as dangerous ([0109]).
or violations of driving rules over the time period.
See at least:
“the predicted behavior of other objects is based not only on the type of object and its current trajectory, but also based on some likelihood that the object may obey traffic rules or pre-determined behaviors.” (Zhu, [0051])
Rationale:
Zhu expressly teaches evaluating whether a tracked road actor obeys traffic rules, and identifies “speeding” (a driving rule violation) as dangerous behavior ([0109]). Recording these violations over the tracking period of [0058] gives violations of driving rules over the time period.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
wherein the plurality of metrics include at least one of:
an average distance between the first road actor and a second road actor in front of the first actor;
an average distance between the first road actor and a lateral road actor;
an average jerk value;
an average delay in deceleration response;
Disclosure by Huang
Huang teaches:
wherein the plurality of metrics include at least one of:
See at least:
“During steady-state vehicle following, the driver's main purpose in headway control is to maintain his or her headway distance or headway time” (Huang, [0130])
Rationale:
Huang expressly teaches that the metrics used for style classification are drawn from distances, speeds, and accelerations of the observed vehicle relative to surrounding vehicles. The specific alternatives are mapped below; because they are recited in the alternative (“at least one of”), any one satisfies the claim. Zhu, as noted above, already teaches two of them.
an average distance between the first road actor and a second road actor in front of the first actor;
See at least:
“the average headway distance, or headway time, the average velocity of the vehicle ... are used as the original features in the classification.” (Huang, [0130])
Rationale:
Huang expressly teaches the average headway distance, i.e., the average distance between a vehicle and the road actor in front of it. In the combination, this metric is computed for the first road actor and its preceding vehicle (a second road actor), using the relative distance Kwon tracks between the observed vehicle and its neighboring vehicle (Kwon [0044]).
an average distance between the first road actor and a lateral road actor;
See at least:
“The range variable R can be estimated as a weighted average between the headway distance ... and the running average of the adjacent lane vehicle gaps” (Huang, [0076])
Rationale:
Huang expressly teaches computing a running average of the gaps to vehicles in adjacent lanes, i.e., an average distance to lateral road actors. Huang uses this average for its traffic index, which is itself an input to style classification ([0096]). Computing the same running average for the first road actor and a lateral road actor is a direct application of Huang's computation to the tracked vehicle.
an average jerk value;
See at least:
“From the lateral motion point of view, the steering angle, yaw rate, lateral acceleration and lateral jerk can all reflect a driver's driving style.” (Huang, [0158])
Rationale:
Huang expressly teaches that jerk reflects driving style, and uses jerk values as style features (e.g., the maximum lateral jerk, [0163]). Huang uses the maximum rather than the average, but Huang summarizes other signals by their averages (“average speed,” [0165]; “average headway distance,” [0130]). Choosing the average rather than the maximum is a choice among a small number of standard statistics that Huang itself uses interchangeably, so an average jerk value would have been obvious to a PHOSITA; the average is the more robust choice against single-sample noise.
an average delay in deceleration response;
See at least:
“When another vehicle cuts in, the subject vehicle usually decelerates until the headway distance/time reaches the steady-state headway distance/time preferred by the driver. A more conservative driver usually decelerates faster to get back to his/her comfort level quicker, while a sportier driver has a higher tolerance of the shorter distance and decelerates relatively slowly.” (Huang, [0134])
Rationale:
Huang expressly teaches using how quickly a driver decelerates in response to a reduced gap, including “the time to reach the preferred headway time” ([0134]), as a driving-style feature. That is a delay in deceleration response. Averaging it over the cut-in events in a time period is the same averaging Huang applies to other features ([0130]), so an average delay in deceleration response would have been obvious to a PHOSITA.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the first road actor's kinematic and relative kinematic parameters collected over the time period are reduced to driving-behavior metrics of the type taught by Huang (for example, average headway distance, average gaps to adjacent-lane vehicles, jerk, and deceleration response, Huang [0076], [0130], [0134], [0158]), and so that the driving style is determined from those metrics. Huang teaches that such metrics are “a better reflection of the driver's driving style” than raw signals (Huang [0130]). Huang also teaches that, because such signals “are time-domain series, data reduction is necessary in order to reduce the complexity of the classifier” (Huang [0131]), and that the resulting style characterization “can be used in various vehicle control algorithms” (Huang [0050]). A PHOSITA would also have recognized that supplying such reduced metrics, rather than raw high-rate signals, to the trained inference model of Kwon ([0056]) gives that model a compact, fixed-length input, which lowers the computation needed for real-time inference on board the vehicle and improves the stability of its output. Kwon already collects the observed vehicle's relative distance and relative speed to neighboring vehicles over time (Kwon [0044], [0067]). Huang's metrics are computed from generic kinematic quantities (range, range rate, speed, acceleration, jerk) that are as available for an observed vehicle as for the host vehicle. Applying them to the road actor that Zhu and Kwon track therefore does not change their principle of operation, and it predictably yields a more stable, lower-complexity style determination. That improves the efficiency, accuracy, and reliability with which a dangerous road actor is identified for Zhu's safety response (Zhu [0109]).
Regarding Claim 17,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 17.
Disclosure by Zhu
Zhu teaches:
generating, using the kinematic parameters and the relative kinematic parameters of the first road actor,
See at least:
“the vehicle's computer 110 also uses the sensor data to determine the object's state, such as speed and lane position. (Block 640).” (Zhu, [0058])
Rationale:
Zhu expressly teaches generating, from the object's speed (a kinematic parameter) and lane position (a relative kinematic parameter), a record of the object's state. In the combination, the relative kinematic parameters with respect to other road actors (Kwon [0044]) are included in that generated record.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
comparing the driving profile of the first road actor to one or more reference driving profiles;
and determining the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
Disclosure by Kwon
Kwon teaches:
a driving profile of the first road actor including a plurality of metrics indicative of driving behaviors of the first road actor;
See at least:
“the driving data may include state information and the action information of the target driving vehicle, and may be observation result data of the target driving vehicle collected in the monitoring unit 110” (Kwon, [0065])
Rationale:
Kwon expressly teaches generating a collection of driving data for the individual target vehicle. It comprises state information (speed, lane, relative distance and relative speed to neighboring vehicles, [0044]) and action information (lane changes and acceleration adjustments, [0044]), each of which indicates how that vehicle is being driven. The collection is a driving profile of the first road actor including a plurality of metrics indicative of its driving behaviors.
Motivation to Combine Zhu and Kwon
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu and Kwon before them, to combine the references as set forth with respect to Claim 11, and further to organize the first road actor's kinematic and relative kinematic parameters, together with its action information, into a driving profile of that individual road actor as taught by Kwon ([0065]). This gives the style determination a single structured input per road actor, which is a predictable arrangement of known data that improves the reliability of the per-actor determination for the reasons set forth with respect to Claim 11.
Claim Limitations Not Explicitly Taught by the Combination of Zhu and Kwon
After combining the teachings of Zhu and Kwon, the following claim limitations are not explicitly taught:
comparing the driving profile of the first road actor to one or more reference driving profiles;
and determining the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
Disclosure by Huang
Huang teaches:
comparing the driving profile of the first road actor to one or more reference driving profiles;
See at least:
“The basic idea of the FCM-based classification is to determine the class of a pattern, which is represented by a discriminant vector, based on its distance to each pre-determined cluster center. ... C is the total number of pre-determined clusters, such as C=3-5 representing the different levels of sporty driving.” (Huang, [0173])
Rationale:
Huang expressly teaches comparing a driving profile (the “discriminant vector” of driving-behavior metrics) to a set of pre-determined cluster centers. Each cluster center is a reference profile representing a defined level of driving style, so the comparison is to one or more reference driving profiles.
and determining the driving style of the first road actor based on the comparison of the driving profile of the first road actor to the one or more reference driving profiles.
See at least:
“Based on the distances, the algorithm further determines the membership degree” (Huang, [0174])
Rationale:
Huang expressly teaches determining the style class from the computed distances to the reference cluster centers, i.e., determining the driving style based on the comparison of the driving profile to the one or more reference driving profiles.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon to determine the first road actor's driving style by comparing its driving profile with pre-determined reference profiles, as taught by Huang's classification against pre-determined cluster centers that represent defined driving-style levels (Huang [0173]-[0174]). Huang expressly teaches that “fuzzy logic, clustering, neural networks (NN), self-organizing maps (SOM), and even simple threshold-base logic” are alternative techniques for classifying driving style (Huang [0103]). Using Huang's reference-profile comparison in place of, or in addition to, Kwon's inference model is therefore a simple substitution of one known style classifier for another, with predictable results. A PHOSITA would have been motivated to make it for three reasons. First, the comparison yields a membership degree in each style class (Huang [0174]), so the system can directly gauge how closely an observed road actor matches a dangerous profile. That improves the transparency and reliability of Zhu's control decision. Second, the reference profiles are “determined during the design phase” (Huang [0173]), which reduces on-board computation. Third, the predefined style classes can be defined to correspond to the dangerous driving Zhu seeks to avoid (Zhu [0109]).
Regarding Claim 19,
The combination of Zhu and Kwon establishes the method of Claim 11, which is the basis for Claim 19.
Disclosure by Zhu
Zhu teaches:
obtaining information indicative of at least one of a road condition or weather condition;
See at least:
“if the vehicle determines that another object is a bicycle that is beginning to ascend a steep hill in front of the vehicle, the computer may predict that the bicycle will soon slow down” (Zhu, [0048])
Rationale:
Zhu expressly teaches obtaining information indicative of a road condition (a steep grade on the road ahead) and using it to evaluate a road actor's behavior. The limitation is recited in the alternative (“at least one of a road condition or weather condition”), so the road-condition alternative satisfies it.
Claim Limitations Not Explicitly Taught by Zhu
Zhu does not explicitly teach the following claim limitations:
and determining the driving style of the first actor further based on the information indicative of the at least one of the road condition or weather condition.
Disclosure by Huang
Huang teaches:
and determining the driving style of the first actor further based on the information indicative of the at least one of the road condition or weather condition.
See at least:
“the style classification processor 52 classifies driving style based on driver inputs and vehicle motion, as well as the traffic/road conditions. That is, the traffic/road condition indexes are part of the discriminant features (discussed below) used in the style classification.” (Huang, [0096])
Rationale:
Huang expressly teaches determining the driving style further based on information indicative of road conditions. Huang's road condition index includes weather (“weather condition, such as fog, rain, snow,” [0089]), so the determination is based on the road condition, the weather condition, or both.
Motivation to Combine Zhu, Kwon, and Huang
Therefore, given the teachings as a whole, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, having Zhu, Kwon, and Huang before them, to modify the system of Zhu as modified by Kwon so that the driving-style determination for the first road actor is further based on road and weather information, as taught by Huang ([0089], [0096]). Huang teaches making “the traffic/road condition indexes ... part of the discriminant features” used in style classification (Huang [0096]), including ambient conditions such as “fog, rain, snow” (Huang [0089]), because conditions determine “whether the current driving behavior is valuable for the characterization” (Huang [0097]). A PHOSITA would have recognized that the same behavior can be ordinary in one condition and dangerous in another; for example, a following distance that is acceptable on dry pavement may be reckless in rain. Conditioning the style determination on such information therefore improves its accuracy and reduces both false and missed identifications of dangerous road actors, improving the safety of Zhu's control response. Zhu already obtains road information (Zhu [0048]), and Huang's indexes are simply additional classifier inputs. The combination is thus a predictable use of known elements according to their established functions.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OLUWABUSAYO ADEBANJO AWORUNSE whose telephone number is (571)272-4311. The examiner can normally be reached M - F (8:30AM - 5PM).
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jelani Smith can be reached at (571) 270-3969. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/OLUWABUSAYO ADEBANJO AWORUNSE/Examiner, Art Unit 3662
/JELANI A SMITH/Supervisory Patent Examiner, Art Unit 3662
