DETAILED ACTION
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 8/21/2026 has been entered.
This office action is in response to the amendment filed on 8/21/2026. In the amendment, claims 1-2, 9 and 13-14 have been amended, claim 8 is now canceled and claims 17-20 are withdrawn. Overall, claims 1-7 and 9-20 are pending in this application.
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 .
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-7 and 9-16 are rejected under 35 U.S.C. 101 because the claimed invention is directed to abstract idea without significantly more.
[101 Analysis Step 1]
Step 1, of the 2019 Guidance, first looks to whether the claimed invention is directed to a statutory category, namely a process, machine, manufactures, and compositions of matter.
The claim 1 is directed to a method (i.e. process) and claim 13 is directed to an apparatus for machine-learning of a vehicular wait event (i.e. machine). Thus, claims 1 and 13 are one of four the statutory categories (Step 1: YES).
[101 Analysis Step 2A, Prong I]
Regarding Prong I of the Step 2A analysis in the 2019 PEG, the claims are to be analyzed to determine whether they recite subject matter that falls within one of the follow groups of abstract ideas: a) mathematical concepts, b) certain methods of organizing human activity, and/or c) mental processes.
Independent Claim 1 includes limitations that recite an abstract idea (emphasized below) and will be used as a representative claim(s) for the remainder of the 101 rejection.
Claim 1 recites:
A method comprising:
processing a sensor observation to determine a wait event, wherein the wait event indicates that at least one person is in a wait state inside a stationary vehicle;
processing the sensor observation to determine one or more contextual features associated with a location of the wait event, a time of the wait event, the at least one person, or a combination thereof;
determining a ground truth of the wait state;
vectorizing the one or more contextual features and the ground truth into a training vector;
using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors, wherein the predicted waiting data includes a predicted reason for the at least one person being in the wait state; and
providing the trained machine learning model as an output.
The examiner submits that the foregoing bolded limitations(s) constitute a “mental process” because under its broadest reasonable interpretations, the claim covers performance of the limitation in the human mind. For example, a) “processing a sensor observation to determine a wait event, wherein the wait event indicates that at least one person is in a wait state inside a stationary vehicle” this is an observation and evaluation such as a person viewing a camera image or other observation data and making a judgement, b) “processing the sensor observation to determine one or more contextual features associated with a location of the wait event, a time of the wait event, the at least one person, or a combination thereof” this is an observation and making a determination, c) “determining a ground truth of the wait state” this is an evaluation and making a judgment as to the wait state is true or not, d) “vectorizing the one or more contextual features and the ground truth into a training vector” this step organizes the observed information and judgments into an ordered list, a human can do this in paper and pen, e) “to determine predicted waiting data based on one or more input vectors, wherein the predicted waiting data includes a predicted reason for the at least one person being in the wait state” this can be a judgement or opinion as to why the person is waiting, therefore in the context of the claim encompasses a person looking at and using the data collected to formulating a judgement. Accordingly, the claim recites at least one abstract idea.
[101 Analysis Step 2A, Prong II]
Regarding Prong II of the Step 2A analysis in the 2019 PEG, the claims are to be analyzed to determine whether the claim, as a whole, integrates the abstract into a practical application. As noted in the 2019 PEG, it must be determined whether any additional elements in the claim beyond the abstract idea integrate the exception into a practical application in a manner that imposes a meaningful limit on the judicial exception. The courts have indicated that additional elements merely using a computer to implement an abstract idea, adding insignificant extra solution activity, or generally linking use of a judicial exception to a particular technological environment or field of use do not integrate a judicial exception into a “practical application.”
In the present case, the additional limitations beyond the above-noted abstract idea are as follows (where the underlined portions are the “additional limitations” while the bolded portions continue to represent the “abstract idea”):
Claim 1 recites:
A method comprising:
processing a sensor observation to determine a wait event, wherein the wait event indicates that at least one person is in a wait state inside a stationary vehicle;
processing the sensor observation to determine one or more contextual features associated with a location of the wait event, a time of the wait event, the at least one person, or a combination thereof;
determining a ground truth of the wait state;
vectorizing the one or more contextual features and the ground truth into a training vector;
using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors, wherein the predicted waiting data includes a predicted reason for the at least one person being in the wait state; and
providing the trained machine learning model as an output.
For the following reason(s), the examiner submits that the above identified additional limitations do not integrate the above-noted abstract into a practical applications.
Regarding the additional limitations of “a sensor observation”, “using the training vector to train a machine learning model” and “providing the trained machine learning model as an output” the examiner submits that these limitations are insignificant extra-solution activities that merely use a computer (processing circuitry of a computer system) to perform the process. In particular, the receiving steps can be performed via sensors are recited at a high level of generality (i.e. as a general means of gathering data), and amounts to mere data gathering, which is a form of insignificant extra-solution activity, and regarding “using the training vector to a machine learning model” only requires the generic machine model be trained with the training vector to produce data therefore the limitation merely uses machine learning as a tool to automate the prediction such as “apply it” (see MPEP 2106.05(f)), and “providing the trained model as an output” is insignificant post-solution actively. Lastly, the “processor”, and “sensor” are recited at a high-level of generality (i.e. as a generic processor performing a generic computer function; and sensing data) such that it amounts no more than mere instructions to apply the exception using a generic computer component and data gathering by the sensor.
Thus, taken alone, the additional elements do not integrate the abstract idea into a practical application. Further, looking at the additional limitation(s) as an ordered combination or as a whole, the limitation(s) add nothing that is not already present when looking at the elements taken individually. For instance, there is no indication that the additional elements, when considered as a whole, reflect an improvement in the functioning of a computer or an improvement to another technology or technical filed, apply or use the above-noted judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition, implement/use the above-noted judicial exception with a particular machine or manufacture that is integral to the claim, effect a transformation or reduction of a particular article to a different state or thing, or apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is not more than drafting effort designed to monopolize the exception (MPEP § 2106.05). Accordingly, the additional limitation(s) do/does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea.
[101 Analysis Step 2B]
Regarding Step 2B of the Revised Guidance, representative independent claims 1 and 13 do not include additional elements (considered both individually and as an ordered combination) that are sufficient to amount to significantly more than the judicial exception for the same reasons to those discussed above with respect to determining that the claim does not integrate the abstract idea into a practical application. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements of using a processing circuitry of a computer system (machine learning) with sensor to perform the using… amounts to nothing more than mere instructions to apply the exception using a generic computer component in generic system. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept and performing the functions of “processing…”, “processing…”, “determining…” and “vectorizing…” by person’s mind. And as discussed above, the additional limitations of “providing…” the examiner submits that these limitations are insignificant extra-solution activities. Hence, the claims are not patent eligible.
Dependent claims 2-7, 9-12 and 14-16 do not recite any further limitations that cause the claims to be directed towards statutory subject matter. The claims merely recite: abstract idea. Each of the further limitations expound upon the abstract ideas and do not recite additional elements integrating the abstract ideas into a practical application or additional elements that are not well-understood, routine or conventional. Therefore, dependent claims 2-7, 9-12 and 14-16 are similarly rejected as being directed towards non-statutory subject matter.
Therefore, claims 1-7 and 9-16 is/are ineligible under 35 USC §101.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 1-6, 9-10 and 12-16 are rejected under 35 U.S.C. 103 as being unpatentable over Pub No. US 2022/0398872 A1 to Udipi et. al. (Udipi) in view of Pub No. US 2024/0020642 A1 to Akamine et. al. (Akamine).
In Reference to Claim 1
Udipi teaches (except for the bolded and italic recitations below):
A method comprising:
processing a sensor observation to determine a wait event, wherein the wait event indicates that at least one person is in a wait state inside a stationary vehicle (see at least Udipi Figs. 1-4 and paragraphs [0014], [0048] and [0051] “One non-limiting example of the present disclosure pertains to automatic generation and rendering of a GUI notification that comprises an activity determination (e.g., vehicle activity determination) predicting why a motor vehicle is stopped, for example, on a highway. Activity signal data, comprising location data and motion activity data, is detected from a mobile computing device. Exemplary motion activity data comprises: signal data of the mobile computing device received from different sensors associated with the mobile computing device including: an accelerometer; a gyroscope; a magnetometer; or a combination thereof. In further examples, the activity signal data comprises signal data indicating a state of operation of the mobile computing device such as whether the mobile computing device is in an idle state. Other types of signal data described herein are also includable in activity signal data. Analysis of the activity signal data yields a determination that the mobile computing device is onboard a vehicle (e.g., motor vehicle). The activity signal data is further analyzed to determine that the vehicle (e.g., motor vehicle) is stopped, for example, on a specific type of road. A trained model (e.g., trained AI model) is applied, where the trained model is adapted to generate confidence scoring for generation of one or more activity determinations predicting why the motor vehicle is stopped. Exemplary confidence scoring is generated based on a contextual analysis that comprises an evaluation of the activity signal data. The contextual analysis derives parameters to generate confidence scoring. For example, confidence scoring is generated based on analysis of two or more of: driving behavior of the motor vehicle prior to the motor vehicle being stopped; a determination as to a type of road, of the specific road, that the motor vehicle is stopped on; an evaluation of traffic information associated with a specific road (e.g., highway) that the motor vehicle is stopped on; an evaluation of historical driving patterns associated with a user of the mobile computing device relative to roads traveled/taken; and an evaluation of historical driving behaviors associated with the user of the mobile computing device, among other types of data” and “During real-time (or near real-time) evaluation, flow of method 200 proceeds to processing operation 206. At processing operation 206, signal data is detected for contextual real-time (or near real-time) evaluation of a vehicle. Non-limiting examples of signal data have been described in the foregoing description. In at least one example, signal data comprises activity signal data detected from a mobile computing device. Exemplary activity signal data comprises location data and motion activity data that is detected from one or more sensors of a mobile computing device that is onboard/traveling with a vehicle. The motion activity data comprises: signal data of the mobile computing device received from different sensors associated with the mobile computing device including: an accelerometer; a gyroscope; a magnetometer; or a combination thereof. In further examples, the activity signal data comprises signal data indicating a state of operation of the mobile computing device such as whether the mobile computing device is in an idle state. Other types of signal data described herein are also includable in activity signal data. Analysis of the activity signal data yields a determination that the mobile computing device is onboard a vehicle (e.g., motor vehicle)” and “Flow of method 200 then proceeds to processing operation 210. At processing operation 210, signal data is further analyzed to determine that the vehicle (e.g., motor vehicle) is stopped, for example, on a specific type of road. Similar to the analysis to determine a vehicle type, signal data, including activity signal data, is comparatively analyzed to determine that a vehicle has come to a stop. For instance, a plurality of points of locational data can be analyzed to determine the amount of distance traveled over a given time period as well as other signal data (e.g., speed, velocity, direction), which can provide a determination as to whether a vehicle has stopped moving. In some alternative examples, processing operation 210 is not required as an activity determination of a vehicle may not be based on whether a vehicle is stopped. However, in other examples, a trained AI model focuses on whether a vehicle is stopped and progresses to determine a reason/rationale as to why the vehicle is stopped. Furthermore, locational data (e.g., GPS coordinates) may be further analyzed to determine a specific type of road that a vehicle is stopped on”;
processing the sensor observation to determine one or more contextual features associated with a location of the wait event, a time of the wait event, the at least one person, or a combination thereof (see at least Udipi Figs. 1-4 and paragraphs [0053] and [0054]“In doing so, a trained model (e.g., trained AI model) is configured to retrieve map data from a knowledge repository (e.g., one or more knowledge repositories 110 of FIG. 1) that provides map data for specific locational data (e.g., GPS coordinates) associated with a mobile computing device. Exemplary map data may be retrieved in either a vector format or a bitmap format. In one example, a trained model is adapted to retrieve map data, associated with GPS coordinates, in a bitmap format. Doing so makes it easier (and more efficient) to analyze location data and subsequently match location data with known roadways such as local roads and/or highways. With a vector format, there may be more to unpack. As such, different versions of a trained model (e.g., trained AI model) may be applied based on latency requirements of applications/services that may be interfacing with components that generate vehicle activity determinations. Initially, processing to evaluate map data may be time and labor intensive, but a corpus of map data may be retrieved and analyzed to effectively enable real-time (or near real-time) processing to seamlessly occur with minimal latency. For instance, map data may be analyzed, where different types of roads may be identified through edge detection processing and subsequently labeled (e.g., color-coded identifications). In further technical instances, specific landmarks, objects, etc. may also be labeled and used as points of reference. For instance, modified map data may identify stop signs, traffic lights, building etc., where such data can be used to aid contextual evaluation. For example, a determination can be made from analysis of modified map data that a user is traveling through a school zone. Continuing that example, analysis of activity signal data could indicate that a user is speeding through the school zone. If it is detected that the user is stopped right outside the school zone, the collective contextual analysis can provide a strong indication that the user may have been pulled over for speeding through the school zone. Other types of signal data can further be analysis to help confirm (or disprove) this predictive determination”, “As such, modified map data is created that is adapted for the purpose of roadway identification at run-time of a trained AI model. Recall of map data at run-time can yield fast and accurate determinations as to the type of roadway that a vehicle may be traveling on as well as aid vehicle activity determinations. At run-time, a quick lookup operation is executed using modified map data that is stored and recalled by a trained model. A trained model is further adapted to interface with mapping data resources to continuously make sure that map data is current and up to date. In some examples, a trained AI model is adapted to generate activity determinations of a vehicle only in contextual situations where a specific type of road/roadway is identified as a location (e.g., where a vehicle is stopped). For instance, if a motor vehicle is stopped on a highway, subsequent contextual analysis is then executed to generate a vehicle activity determination. In other technical instances, a vehicle activity determination is generated for a vehicle traveling on any type of road/roadway”;
determining a ground truth of the wait state (see at least Udipi Figs. 1-4 and paragraphs [0037], [0038] “To adapt AI processing for generation of activity determinations, among other determinations, number processing operations are executed comprising processing operations for: data collection; data labeling; training; hyperparameter tuning; and evaluation/re-training. Such processing is used to generate trained AI modeling that satisfies precision and accuracy thresholds set by developers. Accuracy of trained modeling may vary based on different activity determinations generated. For instance, depending on the complexity of the activity determination (e.g., vehicle is stopped versus determining a reason as to why a vehicle is stopped), developers may set different accuracy thresholds for confirming an activity determination. Thresholds set for specific activity determinations may vary without departing from the spirit of the present disclosure. Moreover, as a trained AI model is a learning model, accuracy can improve over time through iterations of training, receipt of feedback (e.g., user feedback as to accuracy of activity determinations), etc” and “With respect to data collection, telemetric analysis is applied to ingest data to adequately train an AI model. Data ingestion processing for building of trained AI processing is known to one skilled in the field of art. Above what is traditionally known, data collection for a trained AI model focuses on specific data types that build a rich model for generating activity determinations of a vehicle. Importantly, it is to be recognized that signal data is collected, stored, and accessed in compliance with privacy laws and regulations. That is, parameters analyzed by a trained AI model are reflective of privacy laws and regulations. Trained AI processing further generates trained AI models that are privacy compliant with respect to access to user-specific data. In some technical instances, users may consent to having user-specific data accessed to improve their user experience and/or operation of an OS and associated applications/services. In such technical instances, trained AI modeling may incorporate user-specific parameters and/or signal data in its contextual analysis. As indicated in the foregoing description, developers may utilize different versions of trained AI models that can be selectively applied based on different contextual scenarios. Furthermore, exemplary modeling is trained and executed in a privacy compliant manner, where developers may never see/access user-specific data due to compliance requirements/restrictions. Moreover, data is retained in compliance with data retention policies and regulations”);
vectorizing the one or more contextual features and the ground truth into a training vector (see at least Udipi Figs. 1-4 and paragraphs [0040] “An integral part of the training process is application of a training algorithm that adapts a trained AI model for fusion analysis to generate activity determinations of a vehicle. As previously referenced, a trained AI model is trained and adapted to apply a fusion algorithm that evaluates a variety of types of signal data in aggregation, where said different types of signal data are abstracted and correlated, to generate accurate vehicle activity determinations. Analysis of such types of signal data in an aggregate manner may be useful in helping generate contextually relevant determinations, data insights, etc. Analysis of exemplary signal data may comprise identifying correlations and relationships between the different types of signal data, where telemetric analysis may be applied to generate determinations with respect to a contextual state of user activity with respect to one or more computing devices. For instance, a user may prefer to receive notifications through a specific OS GUI menu or host application/service endpoint as compared with other GUI menus and/or host application/service endpoints. As an example, a decision tree (e.g., a Gradient Boosting Decision Tree) is used to train an AI models. Above what is traditionally known, an exemplary training algorithm focuses on the defined list of parameters for generating determinations of activity determinations. In doing so, one or more types of decision tree algorithms can be applied generating any number of decision trees to fine-tune a trained AI model. Some parameters for generation of activity determinations are categorical and others may be continuous numbers. As such, additional processing operations may be applied to convert categorical parameters to a numerical representation including but not limited to: Gini, information gain and/or encoding, among other technical examples. An output result of a trained AI model may be a raw data value that is generated from any combination of types of signal data described above. An exemplary raw data value is a numerical representation that correlates with a confidence score for accuracy of a predictive activity determination. For instance, a raw data value may be 0.8333, where that raw data value can be rounded up (or down in alternative instances) to provide a confidence score (e.g., eighty-three (83) percent)”);
using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors, wherein the predicted waiting data includes a predicted reason for the at least one person being in the wait state (see at least Udipi Figs. 1-4 and paragraphs [0052], [0053] “As previously referenced, a level of curation may be applied to determine that high-accuracy locational data is obtained. As an example, a threshold is set relative to the locational data to determine how accurate the locational data is. If the threshold is satisfied, indicating that the locational data is high-accuracy, subsequent processing is executed to determine a specific road that a user is on. For example, a trained AI model is adapted to generate a determination as to the likelihood that a user is driving (or stopped) on a highway as opposed to a local road. For instance, a trained model is adapted to detect when a motor vehicle is stopped on a highway, as compared with a local road, which may provide a stronger indication that a driver may require assistance. A highway is any road with few stops and high speed limits, where different types of highways (e.g., freeways, interstate highways) are known to one skilled in the field of art, all of which are intended to be covered by the teachings of the present disclosure A freeway is a highway that doesn't have a toll. However, it is to be recognized that a trained model (e.g., trained AI model) is adaptable to generate vehicle activity determinations on any type of road. With respect to specific roads, a comprehensive analysis of signal data, as described herein, can yield accurate and precise determinations with respect to vehicle activity including why a vehicle is stopped. For instance, a vehicle may be stopped at a stop sign, stuck in traffic, picking up a passenger, etc., which can be differentiated from contextual scenarios where a notification (e.g., emergency alert notification) may need to be generated” and “In doing so, a trained model (e.g., trained AI model) is configured to retrieve map data from a knowledge repository (e.g., one or more knowledge repositories 110 of FIG. 1) that provides map data for specific locational data (e.g., GPS coordinates) associated with a mobile computing device. Exemplary map data may be retrieved in either a vector format or a bitmap format. In one example, a trained model is adapted to retrieve map data, associated with GPS coordinates, in a bitmap format. Doing so makes it easier (and more efficient) to analyze location data and subsequently match location data with known roadways such as local roads and/or highways. With a vector format, there may be more to unpack. As such, different versions of a trained model (e.g., trained AI model) may be applied based on latency requirements of applications/services that may be interfacing with components that generate vehicle activity determinations. Initially, processing to evaluate map data may be time and labor intensive, but a corpus of map data may be retrieved and analyzed to effectively enable real-time (or near real-time) processing to seamlessly occur with minimal latency. For instance, map data may be analyzed, where different types of roads may be identified through edge detection processing and subsequently labeled (e.g., color-coded identifications). In further technical instances, specific landmarks, objects, etc. may also be labeled and used as points of reference. For instance, modified map data may identify stop signs, traffic lights, building etc., where such data can be used to aid contextual evaluation. For example, a determination can be made from analysis of modified map data that a user is traveling through a school zone. Continuing that example, analysis of activity signal data could indicate that a user is speeding through the school zone. If it is detected that the user is stopped right outside the school zone, the collective contextual analysis can provide a strong indication that the user may have been pulled over for speeding through the school zone. Other types of signal data can further be analysis to help confirm (or disprove) this predictive determination”); and
providing the trained machine learning model as an output (see at least Udipi Figs. 1-4 and paragraphs [0046], [0047] “Train of AI processing may further comprise generating an AI model that is tuned to reflect specific metrics for accuracy, precision and/or recall before a trained AI model is exposed for real-time (near real-time) usage. Developers may set thresholds for specific analytics and/or metrics to make sure that a trained AI model is operating as expected. Thresholds for metric evaluation of a specific trained AI model may vary, depending on developer specifications, without departing from the spirit of the present disclosure. As an example, thresholds may be set relative to an accuracy of an activity determination of a vehicle, determination of usability of locational data and/or motion activity data, and historical signal data, among other examples describes herein. As indicated in the foregoing description, a plurality of trained AI models may be generated and adapted to execute in specific contextual scenarios pertaining to identification of parameters to evaluate. In some technical instances, one or more types of signal data may be unavailable for collection (e.g., one or more sensors are turned off or inactive). As such, a trained AI model may adjust in real-time (or near real-time) to collect applicable signal data and generate an accurate/precise activity determination” and “Once a threshold (or thresholds) is met for exposing a trained AI model, flow of method 200 proceeds to processing operation 204. At processing operation 204, the trained AI modeling is exposed for real-time (or near real-time) evaluation pertaining to generation of an activity determination of a vehicle (vehicle activity determination)”).
Udipi teaches to train the machine learning and output the model however Udipi do not explicitly teaches (bolded and italic recitations above) as to vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output. However, it is known in the art before the effective filing date of the claimed invention to vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output. For example, Akamine teaches to vectorizing the one or more variables into a training vector; using the training vector to train a machine learning model to determine predicted data based on one or more input vectors; and providing the trained machine learning model as an output. Akamine further teaches that performing such steps provide improve accuracy and efficient prediction of values (see at least Akamine Figs. 1-3, 9, 11, 13 and paragraphs 23-27, 104, 108, 110, 113, 115, 118 and 131). Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Udipi to perform the steps of vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output as taught by Akamine in order to improve accuracy and efficient prediction of values. Further the combination Udipi in view of Akamine is the use of known technique to improve a similar method in the same way, yielding predictable results of encoding and organize the labeled training data (See MPEP 2143(I)).
In Reference to Claim 2
The method of claim 1 (see rejection to claim 1 above), wherein the wait state includes the at least one person waiting inside the stationary vehicle (see at least Udipi Figs. 1-4 and paragraphs 12, 14 and 48).
In Reference to Claim 3
The method of claim 1 (see rejection to claim 1 above), wherein the sensor observation comprises sensor data captured by one or more sensors of a device, a vehicle, an infrastructure element, or a combination thereof associated with or within a field of view of the at least one person (see at least Udipi Figs. 1-4 and paragraphs 12, 14 and 48).
In Reference to Claim 4
The method of claim 1 (see rejection to claim 1 above), further comprising: map-matching location data of the sensor observation to a map link, an offset on the map link, or a combination thereof to determine the location of the wait event (see at least Udipi Figs. 1-4 and paragraphs 51-55).
In Reference to Claim 5
The method of claim 4 (see rejection to claim 4 above), wherein the one or more contextual features further includes one or more attributes of the map link determined from a geographic database (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
In Reference to Claim 6
The method of claim 5 (see rejection to claim 5 above), wherein the one or more attributes include a functional class, a speed limit, a presence of a road sign, a bi-directionality, a number of lanes, a speed category, a distance to a point of interest, a stopping or parking sign, a designated stopping or parking area, or a combination thereof associated with the map link (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
In Reference to Claim 9
The method of claim 1 (see rejection to claim 1 above), wherein the predicted reason includes at least one of: the at least one person waiting to pick up another person; or the at least one person waiting for a point of interest to open (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
In Reference to Claim 10
The method of claim 1 (see rejection to claim 1 above), further comprising: using the trained machine learning model to determine the predicted waiting data; and initiating at least one of: generating navigation routing data, mapping data, or a combination thereof based on the predicted waiting data; recommending at least one other person for meeting up, carpooling, or a combination thereof based on the predicted waiting data; recommending at least one activity, at least one good, at least one service, marketing information, vehicle infotainment option, or a combination thereof based on the predicted waiting data; delivering at least one good, at least one service, or a combination thereof to a location of a wait event based on the predicted waiting data; presenting a warning message based on the predicted waiting data; or predicting an impact on parking, traffic, or a combination thereof based on the predicted waiting data (see at least Udipi Figs. 1-4 and paragraphs 12-13, 47, 57 and 60-62).
In Reference to Claim 12
The method of claim 1 (see rejection to claim 1 above), wherein the wait state indicates that the at least one person is remaining within a predetermined proximity of the location of the wait event until an occurrence of an anticipated event (see at least Udipi Figs. 1-4 and paragraphs 12, 14, and 51).
In Reference to Claim 13
Udipi teaches (except for the bolded and italic recitations below):
An apparatus for machine-learning of a vehicular wait event comprising:
at least one processor (402); and at least one memory (403) including computer program code (405) for one or more programs, the at least one memory (403) and the computer program code (405) configured to, with the at least one processor (402) (see at least Udipi Figs. 1-4 and paragraphs [0084], [0085] “FIG. 4 illustrates a computing system 401 suitable for implementing processing operations described herein related to generation and management of activity determinations of a vehicle and provision of associated GUI notifications, with which aspects of the present disclosure may be practiced. As referenced above, computing system 401 may be configured to implement processing operations of any component described herein including an exemplary vehicle activity management component (106 of FIG. 1). As such, computing system 401 may be configured as a specific purpose computing device that executes specific processing operations to solve the technical problems described herein including those pertaining to generation of notifications of vehicle activity including emergency notifications (e.g., associated with a stopped vehicle). Computing system 401 may be implemented as a single apparatus, system, or device or may be implemented in a distributed manner as multiple apparatuses, systems, or devices. For example, computing system 401 may comprise one or more computing devices that execute processing for applications and/or services over a distributed network to enable execution of processing operations described herein over one or more applications or services. Computing system 401 may comprise a collection of devices executing processing for front-end applications/services, back-end applications/services or a combination thereof. Computing system 401 comprises, but is not limited to, a processing system 402, a storage system 403, software 405, communication interface system 407, and user interface system 409. Processing system 402 is operatively coupled with storage system 403, communication interface system 407, and user interface system 409. Non-limiting examples of computer system 401 comprise but are not limited to: smart phones, laptops, tablets, PDAs, desktop computers, servers, smart computing devices including television devices and wearable computing devices including VR devices and AR devices, e-reader devices, gaming consoles, vehicle computing systems/devices, and conferencing systems, among other non-limiting examples” and “Processing system 402 loads and executes software 405 from storage system 403. Software 405 includes one or more software components (e.g., 406a and 406b) that are configured to enable functionality described herein. In some examples, computing system 401 may be connected to other computing devices (e.g., display device, audio devices, servers, mobile/remote devices, VR devices, AR devices, etc.) to further enable processing operations to be executed. When executed by processing system 402, software 405 directs processing system 402 to operate as described herein for at least the various processes, operational scenarios, and sequences discussed in the foregoing implementations. Computing system 401 may optionally include additional devices, features, or functionality not discussed for purposes of brevity. Computing system 401 may further be utilized to execute system diagram 100 (FIG. 1), method 200 (FIG. 2), and/or the accompanying description of FIGS. 3A-3E”). cause the apparatus to perform at least the following,
process a sensor observation to determine a wait event, wherein the wait event indicates that at least one person is in a wait state inside a stationary vehicle (see at least Udipi Figs. 1-4 and paragraphs [0014], [0048] and [0051] “One non-limiting example of the present disclosure pertains to automatic generation and rendering of a GUI notification that comprises an activity determination (e.g., vehicle activity determination) predicting why a motor vehicle is stopped, for example, on a highway. Activity signal data, comprising location data and motion activity data, is detected from a mobile computing device. Exemplary motion activity data comprises: signal data of the mobile computing device received from different sensors associated with the mobile computing device including: an accelerometer; a gyroscope; a magnetometer; or a combination thereof. In further examples, the activity signal data comprises signal data indicating a state of operation of the mobile computing device such as whether the mobile computing device is in an idle state. Other types of signal data described herein are also includable in activity signal data. Analysis of the activity signal data yields a determination that the mobile computing device is onboard a vehicle (e.g., motor vehicle). The activity signal data is further analyzed to determine that the vehicle (e.g., motor vehicle) is stopped, for example, on a specific type of road. A trained model (e.g., trained AI model) is applied, where the trained model is adapted to generate confidence scoring for generation of one or more activity determinations predicting why the motor vehicle is stopped. Exemplary confidence scoring is generated based on a contextual analysis that comprises an evaluation of the activity signal data. The contextual analysis derives parameters to generate confidence scoring. For example, confidence scoring is generated based on analysis of two or more of: driving behavior of the motor vehicle prior to the motor vehicle being stopped; a determination as to a type of road, of the specific road, that the motor vehicle is stopped on; an evaluation of traffic information associated with a specific road (e.g., highway) that the motor vehicle is stopped on; an evaluation of historical driving patterns associated with a user of the mobile computing device relative to roads traveled/taken; and an evaluation of historical driving behaviors associated with the user of the mobile computing device, among other types of data” and “During real-time (or near real-time) evaluation, flow of method 200 proceeds to processing operation 206. At processing operation 206, signal data is detected for contextual real-time (or near real-time) evaluation of a vehicle. Non-limiting examples of signal data have been described in the foregoing description. In at least one example, signal data comprises activity signal data detected from a mobile computing device. Exemplary activity signal data comprises location data and motion activity data that is detected from one or more sensors of a mobile computing device that is onboard/traveling with a vehicle. The motion activity data comprises: signal data of the mobile computing device received from different sensors associated with the mobile computing device including: an accelerometer; a gyroscope; a magnetometer; or a combination thereof. In further examples, the activity signal data comprises signal data indicating a state of operation of the mobile computing device such as whether the mobile computing device is in an idle state. Other types of signal data described herein are also includable in activity signal data. Analysis of the activity signal data yields a determination that the mobile computing device is onboard a vehicle (e.g., motor vehicle)” and “Flow of method 200 then proceeds to processing operation 210. At processing operation 210, signal data is further analyzed to determine that the vehicle (e.g., motor vehicle) is stopped, for example, on a specific type of road. Similar to the analysis to determine a vehicle type, signal data, including activity signal data, is comparatively analyzed to determine that a vehicle has come to a stop. For instance, a plurality of points of locational data can be analyzed to determine the amount of distance traveled over a given time period as well as other signal data (e.g., speed, velocity, direction), which can provide a determination as to whether a vehicle has stopped moving. In some alternative examples, processing operation 210 is not required as an activity determination of a vehicle may not be based on whether a vehicle is stopped. However, in other examples, a trained AI model focuses on whether a vehicle is stopped and progresses to determine a reason/rationale as to why the vehicle is stopped. Furthermore, locational data (e.g., GPS coordinates) may be further analyzed to determine a specific type of road that a vehicle is stopped on”);
process the sensor observation to determine one or more contextual features associated with a location of the wait event, a time of the wait event, the at least one person, or a combination thereof (see at least Udipi Figs. 1-4 and paragraphs [0053] and [0054]“In doing so, a trained model (e.g., trained AI model) is configured to retrieve map data from a knowledge repository (e.g., one or more knowledge repositories 110 of FIG. 1) that provides map data for specific locational data (e.g., GPS coordinates) associated with a mobile computing device. Exemplary map data may be retrieved in either a vector format or a bitmap format. In one example, a trained model is adapted to retrieve map data, associated with GPS coordinates, in a bitmap format. Doing so makes it easier (and more efficient) to analyze location data and subsequently match location data with known roadways such as local roads and/or highways. With a vector format, there may be more to unpack. As such, different versions of a trained model (e.g., trained AI model) may be applied based on latency requirements of applications/services that may be interfacing with components that generate vehicle activity determinations. Initially, processing to evaluate map data may be time and labor intensive, but a corpus of map data may be retrieved and analyzed to effectively enable real-time (or near real-time) processing to seamlessly occur with minimal latency. For instance, map data may be analyzed, where different types of roads may be identified through edge detection processing and subsequently labeled (e.g., color-coded identifications). In further technical instances, specific landmarks, objects, etc. may also be labeled and used as points of reference. For instance, modified map data may identify stop signs, traffic lights, building etc., where such data can be used to aid contextual evaluation. For example, a determination can be made from analysis of modified map data that a user is traveling through a school zone. Continuing that example, analysis of activity signal data could indicate that a user is speeding through the school zone. If it is detected that the user is stopped right outside the school zone, the collective contextual analysis can provide a strong indication that the user may have been pulled over for speeding through the school zone. Other types of signal data can further be analysis to help confirm (or disprove) this predictive determination”, “As such, modified map data is created that is adapted for the purpose of roadway identification at run-time of a trained AI model. Recall of map data at run-time can yield fast and accurate determinations as to the type of roadway that a vehicle may be traveling on as well as aid vehicle activity determinations. At run-time, a quick lookup operation is executed using modified map data that is stored and recalled by a trained model. A trained model is further adapted to interface with mapping data resources to continuously make sure that map data is current and up to date. In some examples, a trained AI model is adapted to generate activity determinations of a vehicle only in contextual situations where a specific type of road/roadway is identified as a location (e.g., where a vehicle is stopped). For instance, if a motor vehicle is stopped on a highway, subsequent contextual analysis is then executed to generate a vehicle activity determination. In other technical instances, a vehicle activity determination is generated for a vehicle traveling on any type of road/roadway”);
determine a ground truth of the wait state (see at least Udipi Figs. 1-4 and paragraphs [0037], [0038] “To adapt AI processing for generation of activity determinations, among other determinations, number processing operations are executed comprising processing operations for: data collection; data labeling; training; hyperparameter tuning; and evaluation/re-training. Such processing is used to generate trained AI modeling that satisfies precision and accuracy thresholds set by developers. Accuracy of trained modeling may vary based on different activity determinations generated. For instance, depending on the complexity of the activity determination (e.g., vehicle is stopped versus determining a reason as to why a vehicle is stopped), developers may set different accuracy thresholds for confirming an activity determination. Thresholds set for specific activity determinations may vary without departing from the spirit of the present disclosure. Moreover, as a trained AI model is a learning model, accuracy can improve over time through iterations of training, receipt of feedback (e.g., user feedback as to accuracy of activity determinations), etc” and “With respect to data collection, telemetric analysis is applied to ingest data to adequately train an AI model. Data ingestion processing for building of trained AI processing is known to one skilled in the field of art. Above what is traditionally known, data collection for a trained AI model focuses on specific data types that build a rich model for generating activity determinations of a vehicle. Importantly, it is to be recognized that signal data is collected, stored, and accessed in compliance with privacy laws and regulations. That is, parameters analyzed by a trained AI model are reflective of privacy laws and regulations. Trained AI processing further generates trained AI models that are privacy compliant with respect to access to user-specific data. In some technical instances, users may consent to having user-specific data accessed to improve their user experience and/or operation of an OS and associated applications/services. In such technical instances, trained AI modeling may incorporate user-specific parameters and/or signal data in its contextual analysis. As indicated in the foregoing description, developers may utilize different versions of trained AI models that can be selectively applied based on different contextual scenarios. Furthermore, exemplary modeling is trained and executed in a privacy compliant manner, where developers may never see/access user-specific data due to compliance requirements/restrictions. Moreover, data is retained in compliance with data retention policies and regulations”);
vectorize the one or more contextual features and the ground truth into a training vector (see at least Udipi Figs. 1-4 and paragraphs [0040] “An integral part of the training process is application of a training algorithm that adapts a trained AI model for fusion analysis to generate activity determinations of a vehicle. As previously referenced, a trained AI model is trained and adapted to apply a fusion algorithm that evaluates a variety of types of signal data in aggregation, where said different types of signal data are abstracted and correlated, to generate accurate vehicle activity determinations. Analysis of such types of signal data in an aggregate manner may be useful in helping generate contextually relevant determinations, data insights, etc. Analysis of exemplary signal data may comprise identifying correlations and relationships between the different types of signal data, where telemetric analysis may be applied to generate determinations with respect to a contextual state of user activity with respect to one or more computing devices. For instance, a user may prefer to receive notifications through a specific OS GUI menu or host application/service endpoint as compared with other GUI menus and/or host application/service endpoints. As an example, a decision tree (e.g., a Gradient Boosting Decision Tree) is used to train an AI models. Above what is traditionally known, an exemplary training algorithm focuses on the defined list of parameters for generating determinations of activity determinations. In doing so, one or more types of decision tree algorithms can be applied generating any number of decision trees to fine-tune a trained AI model. Some parameters for generation of activity determinations are categorical and others may be continuous numbers. As such, additional processing operations may be applied to convert categorical parameters to a numerical representation including but not limited to: Gini, information gain and/or encoding, among other technical examples. An output result of a trained AI model may be a raw data value that is generated from any combination of types of signal data described above. An exemplary raw data value is a numerical representation that correlates with a confidence score for accuracy of a predictive activity determination. For instance, a raw data value may be 0.8333, where that raw data value can be rounded up (or down in alternative instances) to provide a confidence score (e.g., eighty-three (83) percent)”);
use the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors, wherein the predicted waiting data includes a predicted reason for the at least one person being in the wait state (see at least Udipi Figs. 1-4 and paragraphs [0052], [0053] “As previously referenced, a level of curation may be applied to determine that high-accuracy locational data is obtained. As an example, a threshold is set relative to the locational data to determine how accurate the locational data is. If the threshold is satisfied, indicating that the locational data is high-accuracy, subsequent processing is executed to determine a specific road that a user is on. For example, a trained AI model is adapted to generate a determination as to the likelihood that a user is driving (or stopped) on a highway as opposed to a local road. For instance, a trained model is adapted to detect when a motor vehicle is stopped on a highway, as compared with a local road, which may provide a stronger indication that a driver may require assistance. A highway is any road with few stops and high speed limits, where different types of highways (e.g., freeways, interstate highways) are known to one skilled in the field of art, all of which are intended to be covered by the teachings of the present disclosure A freeway is a highway that doesn't have a toll. However, it is to be recognized that a trained model (e.g., trained AI model) is adaptable to generate vehicle activity determinations on any type of road. With respect to specific roads, a comprehensive analysis of signal data, as described herein, can yield accurate and precise determinations with respect to vehicle activity including why a vehicle is stopped. For instance, a vehicle may be stopped at a stop sign, stuck in traffic, picking up a passenger, etc., which can be differentiated from contextual scenarios where a notification (e.g., emergency alert notification) may need to be generated” and “In doing so, a trained model (e.g., trained AI model) is configured to retrieve map data from a knowledge repository (e.g., one or more knowledge repositories 110 of FIG. 1) that provides map data for specific locational data (e.g., GPS coordinates) associated with a mobile computing device. Exemplary map data may be retrieved in either a vector format or a bitmap format. In one example, a trained model is adapted to retrieve map data, associated with GPS coordinates, in a bitmap format. Doing so makes it easier (and more efficient) to analyze location data and subsequently match location data with known roadways such as local roads and/or highways. With a vector format, there may be more to unpack. As such, different versions of a trained model (e.g., trained AI model) may be applied based on latency requirements of applications/services that may be interfacing with components that generate vehicle activity determinations. Initially, processing to evaluate map data may be time and labor intensive, but a corpus of map data may be retrieved and analyzed to effectively enable real-time (or near real-time) processing to seamlessly occur with minimal latency. For instance, map data may be analyzed, where different types of roads may be identified through edge detection processing and subsequently labeled (e.g., color-coded identifications). In further technical instances, specific landmarks, objects, etc. may also be labeled and used as points of reference. For instance, modified map data may identify stop signs, traffic lights, building etc., where such data can be used to aid contextual evaluation. For example, a determination can be made from analysis of modified map data that a user is traveling through a school zone. Continuing that example, analysis of activity signal data could indicate that a user is speeding through the school zone. If it is detected that the user is stopped right outside the school zone, the collective contextual analysis can provide a strong indication that the user may have been pulled over for speeding through the school zone. Other types of signal data can further be analysis to help confirm (or disprove) this predictive determination”; and
provide the trained machine learning model as an output (see at least Udipi Figs. 1-4 and paragraphs [0046], [0047] “Train of AI processing may further comprise generating an AI model that is tuned to reflect specific metrics for accuracy, precision and/or recall before a trained AI model is exposed for real-time (near real-time) usage. Developers may set thresholds for specific analytics and/or metrics to make sure that a trained AI model is operating as expected. Thresholds for metric evaluation of a specific trained AI model may vary, depending on developer specifications, without departing from the spirit of the present disclosure. As an example, thresholds may be set relative to an accuracy of an activity determination of a vehicle, determination of usability of locational data and/or motion activity data, and historical signal data, among other examples describes herein. As indicated in the foregoing description, a plurality of trained AI models may be generated and adapted to execute in specific contextual scenarios pertaining to identification of parameters to evaluate. In some technical instances, one or more types of signal data may be unavailable for collection (e.g., one or more sensors are turned off or inactive). As such, a trained AI model may adjust in real-time (or near real-time) to collect applicable signal data and generate an accurate/precise activity determination” and “Once a threshold (or thresholds) is met for exposing a trained AI model, flow of method 200 proceeds to processing operation 204. At processing operation 204, the trained AI modeling is exposed for real-time (or near real-time) evaluation pertaining to generation of an activity determination of a vehicle (vehicle activity determination)”).
Udipi teaches to train the machine learning and output the model however Udipi do not explicitly teaches (bolded and italic recitations above) as to vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output. However, it is known in the art before the effective filing date of the claimed invention to vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output. For example, Akamine teaches to vectorizing the one or more variables into a training vector; using the training vector to train a machine learning model to determine predicted data based on one or more input vectors; and providing the trained machine learning model as an output. Akamine further teaches that performing such steps provide improve accuracy and efficient prediction of values (see at least Akamine Figs. 1-3, 9, 11, 13 and paragraphs 23-27, 104, 108, 110, 113, 115, 118 and 131). Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Udipi to perform the steps of vectorizing the one or more contextual features and the ground truth into a training vector; using the training vector to train a machine learning model to determine predicted waiting data based on one or more input vectors; and providing the trained machine learning model as an output as taught by Akamine in order to improve accuracy and efficient prediction of values. Further the combination Udipi in view of Akamine is the use of known technique to improve a similar method in the same way, yielding predictable results of encoding and organize the labeled training data (See MPEP 2143(I)).
In Reference to Claim 14
The apparatus of claim 13 (see rejection to claim 13 above), wherein the wait state includes the at least one person waiting inside the stationary vehicle (see at least Udipi Figs. 1-4 and paragraphs 12, 14 and 48).
In Reference to Claim 15
The apparatus of claim 13 (see rejection to claim 13 above), wherein the apparatus is further caused to: map-match location data of the sensor observation to a map link, an offset on the map link, or a combination thereof to determine the location of the wait event (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
In Reference to Claim 16
The apparatus of claim 15 (see rejection to claim 15 above), wherein the one or more contextual features further includes one or more attributes of the map link determined from a geographic database (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Udipi in view of Akamine and further in view of Pub No. US 2017/0227370 A1 to O’Mahony et. al. (O’Mahony).
In Reference to Claim 7
Udipi in view of Akamine teaches (except for the bolded and italic recitations below):
The method of claim 1 (see rejection to claim 1 above), wherein the predicted waiting data includes a predicted likelihood of at least one subsequent person waiting at a predicted location, a predicted time, or a combination thereof (see at least Udipi Figs. 1-4 and paragraphs 12, 51-55).
Udipi in view of Akamine does not explicitly teaches (bolded and italic recitations above) as to the predicted waiting data includes a predicted likelihood of at least one subsequent person waiting at a predicted location, a predicted time, or a combination thereof. However, it is known in the art before the effective filing date of the claimed invention that the predicted waiting data includes a predicted likelihood of at least one subsequent person waiting at a predicted location, a predicted time, or a combination thereof. For example, O’Mahony teaches that the predicted waiting data includes a predicted likelihood of at least one subsequent person waiting at a predicted location, a predicted time, or a combination thereof. O’Mahony further teaches that performing such function provides estimate the wait time in the area (see at least O’Mahony Figs. 1-7 and paragraphs 5, 7, 31-36). Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Udipi in view of Akamine with the predicted waiting data includes a predicted likelihood of at least one subsequent person waiting at a predicted location, a predicted time, or a combination thereof as taught by O’Mahony in order to predict the likelihood that the person will wait at a given location and time.
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Udipi in view of Akamine and further in view of Pub. No. US 2019/0250626 A1 to Ghafarianzadeh et. al. (Ghafarianzadeh).
In Reference to Claim 11
Udipi teaches (except for the bolded and italic recitations below):
The method of claim 1 (see rejection to claim 1 above), further comprising: using the trained machine learning model to determine the predicted waiting data; and using the predicted waiting data as an input to an autonomous vehicle control system for at least one of: risk calculation; routing; adapting a safety distance at a specific location, a specific time, or a combination thereof; selecting a travel lane; and sharing a ride (see at least Udipi Figs. 1-4 and paragraphs 14, 47, 51 and 57).
Udipi in view of Akamine does not explicitly teaches (bolded and italic recitations above) as to using the predicted waiting data as an input to an autonomous vehicle control system for at least one of: risk calculation; routing; adapting a safety distance at a specific location, a specific time, or a combination thereof; selecting a travel lane; and sharing a ride. However, it is known in the art before the effective filing date of the claimed invention to using the predicted waiting data as an input to an autonomous vehicle control system for at least one of: risk calculation; routing; adapting a safety distance at a specific location, a specific time, or a combination thereof; selecting a travel lane; and sharing a ride. For example, Ghafarianzadeh teaches to using the predicted waiting data as an input to an autonomous vehicle control system for at least one of: risk calculation; routing; adapting a safety distance at a specific location, a specific time, or a combination thereof; selecting a travel lane; and sharing a ride. Ghafarianzadeh further teaches that performing such steps provide efficiency and safety of the vehicles (see at least Ghafarianzadeh Figs. 1-7 and paragraphs 12, 15-16, 18, 21-22, 35, 38, 77, 80-83, 97). Therefore it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the system of Udipi in view of Akamine to perform the steps of using the predicted waiting data as an input to an autonomous vehicle control system for at least one of: risk calculation; routing; adapting a safety distance at a specific location, a specific time, or a combination thereof; selecting a travel lane; and sharing a ride as taught by Ghafarianzadeh in order to provide efficiency and safety of the vehicles.
Response to Arguments
Applicant’s arguments with respect to claim(s) 1-7 and 9-16 have been considered but are moot because the new ground of rejection does not rely on all of the references applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Applicant's arguments filed 1/29/2026 (regarding 101 rejection) have been fully considered but they are not persuasive.
Applicant argues that “Applicants respectfully traverse the rejection. The August 4, 2025 Memorandum from Deputy Commissioner Charles Kim to Technology Centers 2100, 2600, and 3600 regarding "Reminders on evaluation subject matter eligibility under 35 U.S.C. 101" (hereinafter "the Memo") reminds Examiners that the "mental process grouping is not without limits, Examiners are reminded not to expand this grouping in a manner that encompasses claim limitations that cannot practically be performed in the human mind" (the Memo at page 2, Section II.A) and that "Examiners should be careful to distinguish claims that recite an exception (which require further eligibility analysis) from claims that merely involve an exception (which are eligible and do not require further eligibility analysis." The Memo at page 3 specifically points to USPTO example 39 as illustrating an example claim limitation that "merely involve an abstract idea" and "does not recite a judicial exception." Example 39's recites "training the neural network in a first state using the first training set." The Memo at page 3 then contrasts Example 39 with Example 47 which recites the specific training algorithms "backpropagation algorithm" and "gradient descent algorithm." The Memo explains because Example 47 includes these algorithm, the limitations of Example 47 "require specific mathematical calculations," and therefore, recite a judicial exception. Like Example 39, the claimed subject matter in the instant case recites "using the training vector to train a machine learning model" without specifying the specific training algorithm. Consequently, the claimed subject matter does not recite a judicial exception, and no further patent eligibility analysis is needed. For at least these reasons, claims 1 and 13 are do not recite a judicial exception under Step 2A, Prong 1” however the examiner respectfully disagree with the applicant since the step of using the trained vector to train a machine learning model…” requires a generic “machine learning model” to be trained with the training vector which can be done by a human and the limitation uses machine learning model as a tool to automate the prediction such as “apply it” (see MPEP 2106.05(f)). And further the applicant’s specification in paragraph [0074] discloses that the “machine learning model” is a “traditional convolutional neural network”.
Applicant argues that, “Alternatively, under Step 2A Prong 2, the claims are "integrated into a practical application." The claims provide a technical improvement to navigation and autonomous vehicle safety by providing "situational awareness" of stationary vehicles with waiting passengers that current mapping services fail to detect (see, e.g., Specification at paragraphs [0001], [0027]). This improvement to technical fields beyond the abstract idea constitutes "significantly more" than the judicial exception. As stated in the Memo at page 5, "Claims that are determined to improve computer capabilities or improve technology or a technical field support a finding that the claim integrates the judicial exception into a practical application or amounts to significantly more than the judicial exception itself." In the instant case, the claimed subject matter improves mapping and navigation technologies by enabling the detection of vehicles with passengers in a wait state using machine learning. For at least these reasons, claims 1 and 13 are also patent eligible under Step 2A, Prong 2. While the above analysis of the patent eligibility the claims are nowhere near a close call given the guidance provided in the Memo, the Memo further reminds Examiners that "if it is a 'close call' as to whether a claim is eligible, they should only make a rejection when it is more likely than not (i.e., more than 50%) that the claim is ineligible under 35 U.S.C. 101" and that "unpatentability must be established by a preponderance of the evidence." In summary, the claims are patent eligible under Step 2A, Prong 1, and no further analysis is needed. Even if further analysis is applied under Step 2A, Prong 2, the claims integrate any abstract idea into a practical application under Step 2A, Prong 2, or otherwise provides significantly more under Step 2B” and the examiner respectfully disagree with the applicant since the improvement as disclosed in the specification in paragraphs [0027]-[0029] is to the map system and not to the machine learning model it self such that improvement is the availability and accuracy of information about where, when and why people waiting, is improvement to the abstract idea itself and not to the computer functionality, sensor technology or the machine learning model (see MPEP 2106.05 (a)(II)). And further in Recentive Analytics, Inc. v. Fox Corp., 134 F.4th 1205 (Fed. Cir. Apr. 18, 2025), also stated that in order for the machine learning model to be eligible, the improvement is to the underlying ML technology and the current claims do not improve the machine learning model itself and merely updates and iterative training which is not eligible.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Pub No. US 2020/0182638 A1 to Suzuki et. al (Suzuki) teaches automated driving buses or by a driver.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRANDON DONGPA LEE whose telephone number is (571)270-3525. The examiner can normally be reached Monday - Friday, 8:00 am - 5:00 pm.
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, Aniss Chad can be reached at (571) 270-3832. 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.
/BRANDON D LEE/Primary Examiner, Art Unit 3662 10/30/2026