DETAILED ACTION
Status of Claims
This Office action is responsive to communications filed on 2025-12-29. Claim(s) 3, 7-8, 10, 13, 17, and 21-26 was/were cancelled. Claim(s) 27-32 was/were added. Claim(s) 1-2, 4-6, 9, 11-12, 14-16, 18-20, and 27-32 is/are pending and are examined herein.
Claim(s) 1-2, 4-6, 9, 11-12, 14-16, 18-20, and 27-32 is/are rejected under 35 USC 112(b).
Claim(s) 1-2, 4-6, 9, 11-12, 14-16, 18-20, and 27-32 is/are rejected under 35 USC 103.
Notice of Pre-AIA or AIA Status
The present application, filed on or after 2013-03-16, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Regarding objections for informalities and rejections under 35 USC 112, the applicant’s amendments resolve the concerns raised in the previous Office action, but introduce new concerns as described below.
Regarding rejections under 35 USC 103, the applicant’s remarks have been fully considered but they are moot. The prior art mapping has been updated in view of the applicant’s amendments.
Claim Rejections - 35 USC 112(b)
The following is a quotation of 35 USC 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.
The following is a quotation of 35 USC 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim(s) 1-2, 4-6, 9, 11-12, 14-16, 18-20, and 27-32 is/are rejected under 35 USC 112(b) or 35 USC 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 USC 112, the applicant), regards as the invention.
Claims 1, 11, and 20 are indefinite for at least the following reasons:
First, they include two separate recitations of a false classification of a first object, rendering unclear which of these is the intended antecedent of the subsequent recitation of “the false classification of the first object” that appears in the claim.
Second, they recite:
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object, maneuvering the vehicle based on the true classification of the first object; and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object, maneuvering the vehicle based on the false classification of the first object [emphasis added].
As best understood by the examiner in view of the applicant’s remarks accompanying the amended claims, these limitations appear to be based on [specification, 0021]. However, the disclosures there do not describe maneuvering the vehicle based on whether the classification is true or false. In fact, no automated mechanism appears to be described in the specification for determining whether the classification is in fact true or false. Rather, the disclosures of [specification, 0021] merely describe maneuvering the vehicle based on a classification, and it is the consequences of this maneuvering that are described as differing based on whether the classification happens to be true or false. This creates a conflict between the claims and the specification, rendering the claim indefinite since, according to MPEP 2173.03, a claim is “indefinite when a conflict or inconsistency between the claimed subject matter and the specification disclosure renders the scope of the claim uncertain”.
To avoid both of the above issues of indefiniteness, the examiner suggests:
“establishing a minimum distance threshold for allowing a false classification of a first object; in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a distance threshold and establishing the classification of the first object, maneuvering the vehicle based on the
For the purpose of compact prosecution, the claim is interpreted broadly as encompassing at least this interpretation. Dependent claims inherit the rejection. Moreover, claims 27-32 recite similar language as their respective parent claims and are consequently indefinite for the same reasons as described above.
Claims 27, 29, and 31 additionally recite a weak true-positive classification of the first object but it is not clear what it means for a classification to be “weak true-positive”. The phrase “true-positive classification” would be understood by a person of ordinary skill in the art as referring to a classification of an object into a category to which the object actually belongs, but this is a binary condition (either the object was classified into a category to which it actually belongs, or it wasn’t), so it is not clear what it means for a true-positive classification to be “weak” as recited in the claim. As best understood by the examiner, any assessment as to whether a classification is a “weak true-positive” would appear to be subjective (e.g., individuals may differ as to whether a classification of a cat as a dog would be a “false classification” or a “weak true-positive classification” or even, perhaps, both). MPEP 2106.05(b)(IV) indicates that, in the presence of subjective claim language, an “objective standard must be provided in order to allow the public to determine the scope of the claim” and, in the present case, no such objective criterion is provided. For the purpose of compact prosecution, the indefinite limitations of these dependents are interpreted broadly as encompassing at least the interpretation analogous to that described under the parent claim.
Claim Rejections - 35 USC 103
The following is a quotation of 35 USC 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.
Claims 1-2, 4, 9, 11-12, 14, 18-20, 27, 29, and 31 are rejected as being unpatentable over David FERGUSON et al. (US9043069, published 2015-05-26; hereafter “Ferguson”), in view of Gareth JAMES et al. (An Introduction to Statistical Learning, Chapter 8: Tree-Based Methods, published 2013; hereafter “James”), Min SHAO et al. (US7191150, published 2007-03-13; hereafter “Shao”), and Xujie GAO et al. (US20180229587A1, published 2018-08-16; hereafter, “Gao”).
Claim 1
Ferguson discloses:
A method for creating a machine learning model that is reconfigurable, comprising: ([Ferguson, figure 7 and columns 4 and 12-13]: Ferguson teaches a control system 706 in a vehicle [Ferguson, figure 7 and column 12 line 19 through column 13 line 8], which includes machine learning algorithms that train classifiers for objects in the surrounding environment of a vehicle [Ferguson, column 4 lines 48-67]. This control system is the “machine learning model” recited by the claim.)
creating a first parameter model [that includes first feature values obtained during a training process for the machine learning model,] the first parameter model also including a first base classifier used by the machine learning model to classify objects detected by an ultra-sonic system within a vicinity of a vehicle; ([Ferguson, figure 7 and column 4]: Ferguson teaches training/creating classification models, such as decision trees, in order to classify objects in the environment of a vehicle [Ferguson, column 4 lines 48-67]. It also teaches that the classification models receive input data from the vehicle’s sensor system 704 [Ferguson, figure 7 and column 11 line 41 through column 12 line 18], which collects data from a variety of sources, including sonar [Ferguson, column 4 lines 48-67]. This sensor system with sonar capabilities is the “ultra-sonic system” recited by the claim.)
communicating with a controller in the vehicle to [update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model] to classify the objects detected by the ultra-sonic system; ([Ferguson, figure 7 and columns 11-14]: Ferguson teaches a computer system 710 in vehicle, which has a processor and memory, and which transmits data to and receives data from the control system [Ferguson, figure 7 and column 13 line 36 through column 14 line 13]. This computer system having communicative functions is the “controller” recited by the claim. Ferguson also discloses that the control system classifies objects based on data received from the sensor system, as explained above [Ferguson, column 4 lines 48-67, figure 7, column 11 line 41 through column 12 line 18, column 13 line 36 through column 14 line 13].)
maneuvering the vehicle based on the true classification of the first object; … maneuvering the vehicle based on the false classification of the first object. ([Ferguson, column 13]: Ferguson discloses an obstacle avoidance system 750 which can “identify, evaluate, and avoid or otherwise negotiate obstacles in the environment in which the vehicle 700 is located” [Ferguson, figure 7 and column 13 lines 3-6]. This maps to “maneuvering the vehicle based on the true/false classification of the first object” as recited by the claim (as best understood by the examiner in view of the 112(b) rejections).)
While Ferguson discloses the use of decision trees used for object classification, it does not explicitly disclose that these decision trees include “feature values,” and it does not disclose creating a second model whose performance can be compared against the first. It also does explicitly disclose a distance threshold. In other words, Ferguson might not distinctly disclose:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model,
receiving a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier;
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establishing a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
James is in the field of machine learning. Moreover, Ferguson in view of James discloses:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model, ([James, section 8.1.2 and figure 8.6]: Training a decision tree always results in a classifier that makes use of a series of splits based on the values of some subset of the features that appear in the input [James, section 8.1.2]. For example, “Age” and “MaxHR” are examples of features in the decision tree of [James, figure 8.6 “Top”].)
receiving a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier; ([James, section 8.1.2 and figure 8.6]: James teaches training decision trees using different methods on the same training data, and demonstrates that the resulting trees will typically depend on different sets of features [James, section 8.1.2]. For example, [James, figure 8.6 “Bottom Right”] depicts a “pruned” version of [James, figure 8.6 “Top”], and the pruning removes dependence of the decision tree on certain feature variables that appear in the unpruned version, such as “Age.”)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson with the use of decision trees as disclosed in James because “producing multiple trees… can often result in dramatic improvements in prediction accuracy” [James, page 303], thereby resulting in a more effective system.
Ferguson in view of James might not distinctly disclose:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establishing a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Shao is in the field of machine learning. Moreover, Ferguson in view of James and Shao discloses:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model ([Shao, column 9]: Shao teaches comparing the performance of the champion and challenger models, replacing the champion with the challenger if the latter proves itself superior [Shao, column 9 lines 4-42].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James with the champion/challenger strategy of Shao because “a champion/challenger strategy is used to optimize strategy” and improves the overall effectivity of the system and “the use and implementation of champion/challenger systems is well known and will be evident to one of skill in the art” [Shao, column 9 lines 4-42].
Ferguson in view of James and Shao might not distinctly disclose:
establishing a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Gao is in the field of automated vehicles. Moreover, Ferguson in view of James, Shao, and Gao discloses:
establishing a minimum distance threshold for allowing a false classification of a first object; ([Gao, 0046]: Gao discloses that “the sensitivity of the threat assessment algorithm may be increased when external objects are detected within a threshold distance from the host vehicle” so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046; emphasis added; see also, 0033]. The threshold distance of Gao maps to the “minimum distance threshold” of the claim. Since the threshold distance of Gao is used to increase sensitivity (which, by definition, refers to the true positive rate), this threshold distance falls under the broadest reasonable interpretation of being “for allowing a false classification of a first object” as recited by the claim.)
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is within a first distance threshold from the vehicle maps to “detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is further away from the vehicle than the first distance threshold maps to “detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James and Shao with the distance-based adjustment of sensitivity described in Gao because it allows the vehicle to be “more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], thereby resulting in a safer system overall.
Claim 2
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claim(s). It further discloses:
[The method of claim 1, wherein] the first parameter model and the configurable parameter model are designed using a decision tree arrangement. ([James, section 8.1.2 and figure 8.6; Ferguson, column 4]: Both models are already decision trees, as explained above [James, section 8.1.2, figure 8.6; Ferguson, column 4 lines 48-67].)
The same motivation to combine applies.
Claim 4
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claim(s). It further discloses:
[The method of claim 2, wherein] the decision tree arrangement includes one or more split thresholds between different classes of data. ([James, section 8.1.2 and figure 8.6]: Training decision trees when the input includes ordered variables makes use of thresholds to split the data [James, section 8.1.2]. For example, “Age < 52” is a split threshold depicted in [James, figure 8.6 “Top”].)
The same motivation to combine applies.
Claim 9
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claim(s). It further discloses:
[The method of claim 1, wherein] the configurable parameter model is tested by the machine learning model prior to updating the first parameter model with the configurable parameter model. ([Shao, column 9]: As noted above, Shao discloses testing the challenger strategy and using it to replace the champion strategy only when it is determined that its performance surpasses the champion strategy’s performance [Shao, column 9 lines 4-42].)
The same motivation to combine applies.
Claim 27
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claims. It also discloses:
[The method of claim 1 further comprising:] in response to detecting the first object a third distance from the vehicle that is greater than the minimum distance and establishing a weak-true positive classification of the first object, maneuvering the vehicle based on the weak true-positive classification of the first object. ([Ferguson, column 13; Gao, 0033, 0046]: As best understood by the examiner in view of the 112(b) rejections, this claim merely repeats limitations which are already found in the parent claim. Namely, Ferguson discloses an obstacle avoidance system 750 which can “identify, evaluate, and avoid or otherwise negotiate obstacles in the environment in which the vehicle 700 is located” [Ferguson, figure 7 and column 13 lines 3-6], and Gao discloses that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033] so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046].)
The same motivation to combine applies.
Claim 11
Ferguson discloses:
A system for creating a machine learning model that is reconfigurable, comprising: ([Ferguson, figure 7 and columns 4 and 12-13]: Ferguson teaches a control system 706 in a vehicle [Ferguson, figure 7 and column 12 line 19 through column 13 line 8], which includes machine learning algorithms that train classifiers for objects in the surrounding environment of a vehicle [Ferguson, column 4 lines 48-67]. This control system is the “system” recited by the claim.)
a controller configured to: ([Ferguson, figure 7 and columns 13-14]: Ferguson teaches a computer system 710 in vehicle, which has a processor and memory, and which transmits data to and receives data from the control system [Ferguson, figure 7 and column 13 line 36 through column 14 line 13]. This computer system having communicative functions is the “controller” recited by the claim.)
store a first parameter model [that includes first feature values obtained during a training process for the machine learning model,] the first parameter model also including a first base classifier used by the machine learning model to classify objects detected by an ultra-sonic system within a vicinity of a vehicle; ([Ferguson, figure 7 and column 4]: Ferguson teaches training/creating classification models, such as decision trees, in order to classify objects in the environment of a vehicle [Ferguson, column 4 lines 48-67]. It also teaches that the classification models receive input data from the vehicle’s sensor system 704 [Ferguson, figure 7 and column 11 line 41 through column 12 line 18], which collects data from a variety of sources, including sonar [Ferguson, column 4 lines 48-67]. This sensor system with sonar capabilities is the “ultra-sonic system” recited by the claim.)
to classify the objects detected by the ultra-sonic system; ([Ferguson, figure 7 and column 4]: Ferguson teaches training/creating classification models, such as decision trees, in order to classify objects in the environment of a vehicle [Ferguson, column 4 lines 48-67]. It also teaches that the classification models receive input data from the vehicle’s sensor system 704 [Ferguson, figure 7 and column 11 line 41 through column 12 line 18], which collects data from a variety of sources, including sonar [Ferguson, column 4 lines 48-67]. This sensor system with sonar capabilities is the “ultra-sonic system” recited by the claim.)
maneuver the vehicle based on the true classification of the first object; … maneuver the vehicle based on the false classification of the first object. ([Ferguson, column 13]: Ferguson discloses an obstacle avoidance system 750 which can “identify, evaluate, and avoid or otherwise negotiate obstacles in the environment in which the vehicle 700 is located” [Ferguson, figure 7 and column 13 lines 3-6]. This maps to “maneuvering the vehicle based on the true/false classification of the first object” as recited by the claim as best understood by the examiner in view of the 112(b) rejections.)
While Ferguson discloses the use of decision trees used for object classification, it does not explicitly disclose that these decision trees include “feature values,” and it does not disclose creating a second model whose performance can be compared against the first. It also does explicitly disclose a distance threshold. In other words, Ferguson might not distinctly disclose:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model,
receive a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier;
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
James is in the field of machine learning. Moreover, Ferguson in view of James discloses:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model, ([James, section 8.1.2 and figure 8.6]: Training a decision tree always results in a classifier that makes use of a series of splits based on the values of some subset of the features that appear in the input [James, section 8.1.2]. For example, “Age” and “MaxHR” are examples of features in the decision tree of [James, figure 8.6 “Top”].)
receive a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier; ([James, section 8.1.2 and figure 8.6]: James teaches training decision trees using different methods on the same training data, and demonstrates that the resulting trees will typically depend on different sets of features [James, section 8.1.2]. For example, [James, figure 8.6 “Bottom Right”] depicts a “pruned” version of [James, figure 8.6 “Top”], and the pruning removes dependence of the decision tree on certain feature variables that appear in the unpruned version, such as “Age.”)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson with the use of decision trees as disclosed in James because “producing multiple trees… can often result in dramatic improvements in prediction accuracy” [James, page 303], thereby resulting in a more effective system.
Ferguson in view of James might not distinctly disclose:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Shao is in the field of machine learning. Moreover, Ferguson in view of James and Shao discloses:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model ([Shao, column 9]: Shao teaches comparing the performance of the champion and challenger models, replacing the champion with the challenger if the latter proves itself superior [Shao, column 9 lines 4-42].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James with the champion/challenger strategy of Shao because “a champion/challenger strategy is used to optimize strategy” and improves the overall effectivity of the system and “the use and implementation of champion/challenger systems is well known and will be evident to one of skill in the art” [Shao, column 9 lines 4-42].
Ferguson in view of James and Shao might not distinctly disclose:
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Gao is in the field of automated vehicles. Moreover, Ferguson in view of James, Shao, and Gao discloses:
establish a minimum distance threshold for allowing a false classification of a first object; ([Gao, 0046]: Gao discloses that “the sensitivity of the threat assessment algorithm may be increased when external objects are detected within a threshold distance from the host vehicle” so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046; emphasis added; see also, 0033]. The threshold distance of Gao maps to the “minimum distance threshold” of the claim. Since the threshold distance of Gao is used to increase sensitivity (which, by definition, refers to the true positive rate), this threshold distance falls under the broadest reasonable interpretation of being “for allowing a false classification of a first object” as recited by the claim.)
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is within a first distance threshold from the vehicle maps to “detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is further away from the vehicle than the first distance threshold maps to “detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James and Shao with the distance-based adjustment of sensitivity described in Gao because it allows the vehicle to be “more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], thereby resulting in a safer system overall.
Claims 12, 14, 18, and 29 inherit limitations from claim 11 and recite additional limitations which are substantially similar to claims 2, 4, 9, and 27, respectively, so they are rejected by the same rationale.
Claim 19
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claims. It also discloses:
wherein the fixed parameter model includes static values, ([Ferguson, column 4; James, section 8.1.2]: Ferguson discloses the use of decision trees [Ferguson, column 4 lines 48-67], which necessarily contain “values,” such as the features used for splitting or the thresholds for the splits [James, section 8.1.2].)
and the configurable parameter model is used to update the static values. ([Shao, column 9]: Shao discloses replacing/updating the values of the champion model with those of challenger model [Shao, column 9 lines 4-42].)
The same motivation to combine applies.
Claim 20
Ferguson discloses:
A non-transitory computer-readable medium operable to creating a machine learning model, the non-transitory computer readable medium having computer-readable instructions stored thereon that are operable to perform the following: ([Ferguson, figure 7 and columns 13-14]: Ferguson teaches a computer system 710 in vehicle, which has a processor and memory, and which transmits data to and receives data from a control system 706 [Ferguson, figure 7 and column 13 line 36 through column 14 line 13]. The memory of this computer system is the “non-transitory computer-readable medium” recited by the claim.)
store a first parameter model [that includes first feature values obtained during a training process for the machine learning model,] the first parameter model also including a first base classifier used by the machine learning model to classify objects detected by an ultra-sonic system within a vicinity of a vehicle; ([Ferguson, figure 7 and column 4]: Ferguson teaches training/creating classification models, such as decision trees, in order to classify objects in the environment of a vehicle [Ferguson, column 4 lines 48-67]. It also teaches that the classification models receive input data from the vehicle’s sensor system 704 [Ferguson, figure 7 and column 11 line 41 through column 12 line 18], which collects data from a variety of sources, including sonar [Ferguson, column 4 lines 48-67]. This sensor system with sonar capabilities is the “ultra-sonic system” recited by the claim.)
to classify the objects detected by the ultra-sonic system; ([Ferguson, figure 7 and column 4]: Ferguson teaches training/creating classification models, such as decision trees, in order to classify objects in the environment of a vehicle [Ferguson, column 4 lines 48-67]. It also teaches that the classification models receive input data from the vehicle’s sensor system 704 [Ferguson, figure 7 and column 11 line 41 through column 12 line 18], which collects data from a variety of sources, including sonar [Ferguson, column 4 lines 48-67]. This sensor system with sonar capabilities is the “ultra-sonic system” recited by the claim.)
maneuver the vehicle based on the true classification of the first object; … maneuver the vehicle based on the false classification of the first object. ([Ferguson, column 13]: Ferguson discloses an obstacle avoidance system 750 which can “identify, evaluate, and avoid or otherwise negotiate obstacles in the environment in which the vehicle 700 is located” [Ferguson, figure 7 and column 13 lines 3-6]. This maps to “maneuvering the vehicle based on the true/false classification of the first object” as recited by the claim as best understood by the examiner in view of the 112(b) rejections.)
While Ferguson discloses the use of decision trees used for object classification, it does not explicitly disclose that these decision trees include “feature values,” and it does not disclose creating a second model whose performance can be compared against the first. It also does explicitly disclose a distance threshold. In other words, Ferguson might not distinctly disclose:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model,
receive a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier;
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
James is in the field of machine learning. Moreover, Ferguson in view of James discloses:
[a first parameter model] that includes first feature values obtained during a training process for the machine learning model, ([James, section 8.1.2 and figure 8.6]: Training a decision tree always results in a classifier that makes use of a series of splits based on the values of some subset of the features that appear in the input [James, section 8.1.2]. For example, “Age” and “MaxHR” are examples of features in the decision tree of [James, figure 8.6 “Top”].)
receive a configurable parameter model that includes configured feature values that are different from the first feature values, the configurable parameter model including a modified base classifier; ([James, section 8.1.2 and figure 8.6]: James teaches training decision trees using different methods on the same training data, and demonstrates that the resulting trees will typically depend on different sets of features [James, section 8.1.2]. For example, [James, figure 8.6 “Bottom Right”] depicts a “pruned” version of [James, figure 8.6 “Top”], and the pruning removes dependence of the decision tree on certain feature variables that appear in the unpruned version, such as “Age.”)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson with the use of decision trees as disclosed in James because “producing multiple trees… can often result in dramatic improvements in prediction accuracy” [James, page 303], thereby resulting in a more effective system.
Ferguson in view of James might not distinctly disclose:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Shao is in the field of machine learning. Moreover, Ferguson in view of James and Shao discloses:
update the first parameter model with the configurable parameter model, wherein the machine learning model is updated to use the configurable parameter model ([Shao, column 9]: Shao teaches comparing the performance of the champion and challenger models, replacing the champion with the challenger if the latter proves itself superior [Shao, column 9 lines 4-42].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James with the champion/challenger strategy of Shao because “a champion/challenger strategy is used to optimize strategy” and improves the overall effectivity of the system and “the use and implementation of champion/challenger systems is well known and will be evident to one of skill in the art” [Shao, column 9 lines 4-42].
Ferguson in view of James and Shao might not distinctly disclose:
establish a minimum distance threshold for allowing a false classification of a first object;
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object,
and in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object,
Gao is in the field of automated vehicles. Moreover, Ferguson in view of James, Shao, and Gao discloses:
establish a minimum distance threshold for allowing a false classification of a first object; ([Gao, 0046]: Gao discloses that “the sensitivity of the threat assessment algorithm may be increased when external objects are detected within a threshold distance from the host vehicle” so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046; emphasis added; see also, 0033]. The threshold distance of Gao maps to the “minimum distance threshold” of the claim. Since the threshold distance of Gao is used to increase sensitivity (which, by definition, refers to the true positive rate), this threshold distance falls under the broadest reasonable interpretation of being “for allowing a false classification of a first object” as recited by the claim.)
in response to detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold and establishing a true classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is within a first distance threshold from the vehicle maps to “detecting the first object at a first distance from the vehicle that is less than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
in response to detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold and establishing a false classification of the first object, ([Gao, 0033]: Gao indicates that “the algorithm may be highly sensitive to speed and trajectory of movement when an object is within a first distance threshold from the vehicle. In contrast, similar speed and trajectory of movement may have little or no impact to the risk score at distances further away from the vehicle” [Gao, 0033; emphasis added]. Verifying that an object that is further away from the vehicle than the first distance threshold maps to “detecting the first object at a second distance from the vehicle that is greater than the minimum distance threshold” of the claim. The examiner reiterates that that this determination is used so that so that “the vehicle may be more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], i.e., for “maneuvering” the vehicle.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James and Shao with the distance-based adjustment of sensitivity described in Gao because it allows the vehicle to be “more responsive to more subtle behaviors of external objects when they are close by” [Gao, 0046], thereby resulting in a safer system overall.
Claims 31 inherits limitations from claim 20 and recites additional limitations which are substantially similar to claim 27, so it is rejected by the same rationale.
Claims 5 and 15 are rejected as being unpatentable over Ferguson in view of James, Shao, and Gao, further in view of Yu-cherng WU (US7437703B2, published 2008-10-14; hereafter, “Wu”).
Claim 5
Ferguson in view of James and Shao discloses the elements of the parent claim(s). It does not distinctly disclose invalid values. In other words, it does not distinctly disclose:
[The method of claim 2, wherein] the decision tree arrangement includes one or more invalid value assignments.
Wu is in the field of machine learning. Moreover, Ferguson in view of James, Shao, Gao, and Wu discloses:
[The method of claim 2, wherein] the decision tree arrangement includes one or more invalid value assignments. ([Wu, figure 3, columns 4 and 7]: Wu discloses a system in which the training/arrangement of decision trees, which occurs as a part of a knowledge discovery engine 68 [Wu, column 4 lines 50-62], is preceded [Wu, figure 3] by the processing of outlier/invalid data by the pattern recognition service 56 [Wu, column 7 line 20 through column 8 line 9].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection, classification, and avoidance system of Ferguson in view of James, Shao, and Gao with the data cleansing techniques of Wu because data cleansing allows the model to “recognize meaningful patterns” [Wu, column 7 lines 29-30], thereby resulting in a more effective classification model.
Claims 15 inherits limitations from claim 11 and recite additional limitations which are substantially similar to claim 5, so it is rejected by the same rationale.
Claims 6 and 16 are rejected as being unpatentable over Ferguson in view of James, Shao, and Gao, further in view of Leo BREIMAN et al. (Classification and Regression Trees, published 1984; hereafter, “Breiman”).
Claim 6
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claim(s). It does not distinctly disclose missing values. In other words, it does not distinctly disclose:
[The method of claim 2, wherein] the decision tree arrangement includes one or more missing value assignments.
Breiman is in the field of machine learning. Moreover, Ferguson in view of James, Shao, Gao, and Breiman discloses:
[The method of claim 2, wherein] the decision tree arrangement includes one or more missing value assignments. ([Breiman, section 5.3.2]: Breiman discloses the training/arrangement of decision trees in presence of missing values [Breiman, section 5.3.2].)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection, classification, and avoidance system of Ferguson in view of James, Shao, and Gao with the missing data algorithm of Breiman because it allows one “to make maximum use of the data cases, complete or not, in the tree construction” and “to construct a tree that will classify any case dropped into it, even if the case has some variable values missing” [Breiman, page 142], thereby resulting in a more effective classification model.
Claims 16 inherits limitations from claim 11 and recite additional limitations which are substantially similar to claim 6, so it is rejected by the same rationale.
Claims 28, 30, and 32 are rejected as being unpatentable over Ferguson in view of James, Shao, and Gao, further in view of Christopher OESTERLING et al. (US20220222475A1, effectively filed 2021-01-13; hereafter, “Oesterling”).
Claim 28
Ferguson in view of James, Shao, and Gao discloses the elements of the parent claims. It also discloses:
[The method of claim 1 further comprising:] establishing a second minimum distance threshold for allowing a false classification of a second object, ([Gao, 0033 and 0046]: This limitation is similar to limitations recited in the parent claim and is disclosed in the same way.)
in response to detecting the second object at a third distance from the vehicle that is less than the second minimum distance threshold and establishing a true classification of the second object, maneuvering the vehicle based on the true classification of the second object; ([Ferguson, column 13; Gao, 0033, 0046]: This limitation is similar to limitations recited in the parent claim and is disclosed in the same way.)
and in response to detecting the second object at a fourth distance from the vehicle that is greater than the second minimum distance threshold and establishing a false classification of the second object, maneuvering the vehicle based on the false classification of the second object. ([Ferguson, column 13; Gao, 0033, 0046]: This limitation is similar to limitations recited in the parent claim and is disclosed in the same way.)
Ferguson in view of James, Shao, and Gao might not distinctly disclose:
wherein the second minimum distance threshold is different from the minimum distance threshold;
Oesterling is in the field of automated vehicles. Moreover, Ferguson in view of James, Shao, Gao, and Oesterling discloses:
wherein the second minimum distance threshold is different from the minimum distance threshold; ([Oesterling, 0066]: Oesterling discloses “changing proximity thresholds used in detecting collision threats” [Oesterling, 0066]. In the combination, a changed distance/proximity threshold maps to the “second minimum distance threshold” of the claim, so that “the second minimum distance threshold is different from the minimum distance threshold” as recited in the claim.)
Before the effective filing date of the invention, it would have been obvious to a person of ordinary skill in the art to combine the vehicle-based object detection and classification system of Ferguson in view of James, Shao, and Gao with changing proximity thresholds as described in Oesterling because it allows “more readily output[ting] an indication of a proximal collision threat” [Oesterling, 0066], thereby resulting in a safer system overall.
Claims 30 and 32 inherit limitations from claims 11 and 20, respectively, and recite additional limitations which are substantially similar to claim 28, so they are rejected by the same rationale.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Hamid GOLGIRI et al. (US20190230471A1, published 2019-07-25; hereafter, “Golgiri”) discloses “establish[ing] multiple proximity zones around the vehicle based on distance thresholds” [Golgiri, 0015].
Zhenhua ZHANG et al. (US20220084398A1, effectively filed 2020-11-11; hereafter, “Zhang”) discloses “vary[ing] the distance threshold d1, to obtain different levels of accuracy and coverage” [Zhang, 0077].
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 Shishir AGRAWAL whose telephone number is +1 703-756-1183. The examiner can normally be reached Monday through Thursday, 08:30-14:30 Pacific Time.
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, Alexey SHMATOV can be reached on +1 571-270-3428. The fax phone number for the organization where this application or proceeding is assigned is +1 571-273-8300.
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 +1 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call +1 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/S.A./Examiner, Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123