DETAILED ACTION
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 .
Status of Claims
This communication is in response to application 18/055,990 filed on 12/16/2025. Claims 1, 10 and 16 have been amended. Claims 1-21 are pending and examiner in the instant office action. The rejections are as stated below.
Response to Arguments
Applicant’s arguments submitted on 12/16/2025, with respect to the previous 35 U.S.C. 112(b) rejection has been fully considered. Applicant has amended claims 1,10 and 16, thereby rendering previous rejection moot.
Applicant’s arguments submitted on 12/16/2025, with respect to the previous 35 U.S.C. 101 rejection has been fully considered and unpersuasive.
With respect to the previous 35 U.S.C. 101 rejection of claim 1, Applicant argues the claims are directed to a practical application that integrates technical improvements into aircraft safety systems rather than merely abstract data processing.
Examiner respectfully disagrees. The same updated analysis based on the new 2019 Patent Eligibility Guidance (2019 PEG) applies to the newly added claimed limitations as discussed in the previous office action.
As a result, Step 2A Prong 1 determines if a claim is directed to those grouping and subgroupings along with an explanation of why it is directed to such.
“First, the rejection should identify the judicial exception (i.e., abstract idea enumerated in Section I of the 2019 PEG, laws of nature, or a natural phenomenon) by referring to what is recited (i.e., set forth or described) in the claim and explaining why it is considered to be an exception (Step 2A Prong One). There is no requirement for the examiner to provide further support, such as publications or an affidavit or declaration under 37 CFR 1.104(d)(2), for the conclusion that a claim recites a judicial exception.”
“For abstract ideas, the rejection should explain why a specific limitation(s) recited in the claim falls within one of the enumerated groupings of abstract ideas (i.e., mathematical concepts, mental processes, or certain methods of organizing human activity) or provide a justification for why a specific limitation(s) recited in the claim is being treated as an abstract idea if it does not fall within the enumerated groupings of abstract ideas in accordance with the “tentative abstract idea” procedure in the 2019 PEG.”
In the Non-Final mailed 09/18/2025 examiner performs the analysis and clarifies that “the abstract idea noted in the independent claims…are directed to a “Mental Processes.” Hence, examiner has indicated that these identified limitations are directed to a mental process and has provided a justification for why these limitations fall within one of the enumerated groupings of abstract ideas (i.e. concepts performed in the human mind). This is sufficient under the guidelines of the 2019 PEG and October 2019 Update as cited above. Accordingly, it seems reasonable for the examiner to group the abstract idea under “Mental processes.” as enumerated in Section I of the 2019 PEG.
Prong Two:
With respect to Step 2A, prong two, Integration into a practical application requires an additional element(s) or a combination of additional elements in the claim to apply, rely on, or use the judicial exception in a manner that imposes a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize the exception. Limitations that are not indicative of integration into a practical application are those that are mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea.-see MPEP 2106.05(f).
In contrast, the instant claims are different, the focus of the claims is not on such an improvement in computers as tools, but on certain independently abstract ideas that use computers as tools. The claims here are not directed to a specific improvement to computer functionality nor an inventive solution to any computer specific problem. The claims do not recite an improvement to the computer architecture.
Examiner submits that under the current 35 USC 101 examining practice, the existence of such novel features would still not cure the deficiencies with respect to the abstract idea. See for example: Ultramercial, Inc. v. Hulu, LLC, 112 USPQ2d 1750, U.S. Court of Appeals Federal Circuit, No. 2010-1544, Decided November 14, 2014, 2014 BL 320546, 772 F.3d 709, Page 1754 last two ¶ : Indeed, in this in instant case, the limitations simply narrow or limit the abstract idea without providing anything significantly more than the abstract idea itself.
Lastly, dependent claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements are simply steps performed by a generic computer. The claim merely amounts to the application or instructions to apply the abstract idea on a processor, and is considered to amount to nothing more than requiring a generic processor to merely carry out the abstract idea itself.
Examiner notes the same arguments apply to independent claims 10 and 16
For these reasons the rejection under 35 U.S.C. § 101 directed to non-statutory subject matter set forth in this office action is maintained.
Applicant’s arguments submitted on 12/16/2025, with respect to the previous 35 U.S.C. 103 rejection has been fully considered and unpersuasive.
With respect to the previous 35 U.S.C. 103 rejection of claim 1, Applicant argues that the cited art of record Liu is non-analogous because it does not explicitly relate to aircraft, aircraft trajectories, or runway approach flight paths, and asserts that the rejection relies on impermissible hindsight reconstruction.
Examiner respectfully disagrees. The rejection does not rely on Liu to disclose aircraft-specific subject matter. Rather, Liu is cited for its teaching of a game-theoretic model used to select or optimize among multiple machine-learning or decision models for anomaly prediction and handling. Liu expressly teaches this functional concept. ChenChu discloses receiving position data defining a trajectory and processing such data using learning models to detect anomalies. Liu teaches that incorporating a game-theoretic framework into predictive and anomaly-handling systems improves model selection and decision-making reliability. One of ordinary skill in the art would have been motivated to apply the game-theoretic model selection technique of Liu to the anomaly detection framework of ChenChu in order to improve robustness and reliability of anomaly determination. Applicant’s argument improperly requires the secondary reference to disclose the same field of use or all claim elements. However, a reference need only be analogues with respect to the problem addressed. Here both references address improving anomaly prediction and decision making using machine learning model, and the combination merely applies a known optimization technique to a known anomaly detection system using predictable results.
Examiner notes the same rationale applies to independent claims 10 and 16.
For at least the above, the previous 35 U.S.C. 103 rejection is maintained.
Claim Rejections – 35 USC §101
35 USC §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 1-21 are rejected under 35 USC §101 because the claimed invention is directed to an abstract idea without significantly more. See MPEP 2106 (III).
The determination of whether a claim recites patent ineligible subject matter is a two-step inquiry.
STEP 1: the claim does not fall within one of the four statutory categories of invention (process, machine, manufacture or composition of matter), See MPEP 2106.03, or
STEP 2: the claim recites a judicial exception, e.g. an abstract idea, without reciting additional elements that amount to significantly more than the judicial exception, as determined using the following analysis: See MPEP 2106.04
STEP 2A (PRONG 1): Does the claim recite an abstract idea, law of nature, or natural phenomenon? See MPEP 2106.04(II)(A)(1)
STEP 2A (PRONG 2): Does the claim recite additional elements that integrate the judicial exception into a practical application? See MPEP 2106.04(II)(A)(2)
STEP 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception? See MPEP 2106.05
Claim 1. An anomaly detection device for an aircraft comprising:
a memory and a processor configured to [applying the abstract idea using generic computer components];
receive satellite position data collected by the aircraft and comprising a sequence of aircraft positions defining an aircraft trajectory [pre-solution activity (data gathering)];
process the satellite position data [mental process/step] using a plurality of different deep learning models [applying the abstract idea using generic computing model] to determine respective aircraft trajectory anomalies in a runway approach flight path for the aircraft [mental process/step],
select a given deep learning model that determines the aircraft trajectory anomalies in the runway approach flight path [mental process/step based upon a game theoretic model [applying the abstract idea using generic computing model],
generate an alert if the respective aircraft trajectory anomaly in the runway approach flight path determined by the given deep learning model exceeds a threshold [insignificant post-solution activity (displaying results)].
101 Analysis - Step 1: Statutory category – Yes
The claim recites a device including at least one step. The claim falls within one of the four statutory categories. See MPEP 2106.03.
Step 2A Prong one evaluation: Judicial Exception – Yes – Mental processes
In Step 2A, Prong one of the 2019 Patent Eligibility Guidance (PEG), a claim is to be analyzed to determine whether it recites subject matter that falls within one of the following groups of abstract ideas: a) mathematical concepts, b) mental processes, and/or c) certain methods of organizing human activity.
The Office submits that the foregoing bolded limitation(s) constitutes judicial exceptions in terms of “mental processes” because under its broadest reasonable interpretation, the limitations can be “performed in the human mind, or by a human using a pen and paper”. See MPEP 2106.04(a)(2)(III)
The claim recites the limitation/steps of process the satellite position data to determine respective aircraft trajectory anomalies in a runway approach flight path for the aircraft and select a given deep learning model that determines the aircraft trajectory anomalies in the runway approach flight path. This limitation, as drafted, are simple processes that, under its Broadest Reasonable Interpretation (BRI), covers performance of the limitation in the mind but for the recitation of the “a memory/processor and deep learning models” and the “a memory/processor and deep learning models” in base claim 1. That is, other than reciting “a memory/processor and deep learning models” nothing in the claim elements precludes the steps from practically being performed in the mind. For example, but for a memory/processor and deep learning models language, the claim encompasses a person looking at data collected and forming a simple judgement. The mere nominal recitation by a computer does not take the claim limitations out of the mental process grouping. Thus, the claim recites a mental process.
Step 2A Prong two evaluation: Practical Application - No
In Step 2A, Prong two of the 2019 PEG, a claim is to be evaluated whether, as a whole, it integrates the recited judicial exception into a practical application. As noted in MPEP 2106.04(d), 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, such that the claim is more than a drafting effort designed to monopolize the judicial exception. The courts have indicated that additional elements such as: 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.”
The Office submits that the foregoing underlined limitation(s) recite additional elements that do not integrate the recited judicial exception into a practical application.
The claim recites additional elements or steps of memory, processor and deep learning models. In particular, memory, processor and deep learning models limitations are recited at a high level of generality (i.e. generic processor performing a generic computer function) such that it amounts to no more than mere instructions to “apply” the exception using a generic computer component.
Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Step 2B evaluation: Inventive concept - No
In Step 2B of the 2019 PEG, a claim is to be evaluated as to whether the claim, as a whole, amounts to significantly more than the recited exception, i.e., whether any additional element, or combination of additional elements, adds an inventive concept to the claim. See MPEP 2106.05.
As discussed with respect to Step 2A Prong Two, the additional elements in the claim amount to no more than mere instructions to apply the exception using a generic computer component. The same analysis applies here in 2B, i.e., mere instructions to apply an exception on a generic computer cannot integrate a judicial exception into a practical application at Step 2A or provide an inventive concept in Step 2B. See MPEP 2106.05(f).
Under the 2019 PEG, a conclusion that an additional element is insignificant extra- solution activity in Step 2A should be re-evaluated in Step 2B. Here, the memory/processor and deep learning models were considered to be insignificant extra-solution activity in Step 2A, and thus they are re-evaluated in Step 2B to determine if they are more than what is well-understood, routine, conventional activity in the field.
The Specification does not provide any indication that the “memory/processor and deep learning models” are anything other than a conventional computer within an aircraft (See at least ¶35 of applicant’s specification).
Accordingly, a conclusion that the “memory/processor and deep learning models” elements are well-understood, routine, conventional activity is supported under Berkheimer. Thus, the claim is ineligible.
Independent method claim 10, respectively, recites similar limitations performed by the device of claim 1. Therefore, claim 10 is rejected under the same rationales used in the rejections of claim 1 as outlined above.
Independent medium claim 16, respectively, recites similar limitations performed by the device of claim 1. Therefore, claim 16 is rejected under the same rationales used in the rejections of claim 1 as outlined above.
Dependent claims 2-9, 11-15 and 17-21 do not recite any further limitations that cause the claim(s) to be patent eligible. Rather, the limitations of dependent claims are directed toward additional aspects of the judicial exception and/or well-understood, routine and conventional additional elements that do not integrate the judicial exception into a practical application and amounts to mere input and/or output data manipulation. Therefore, dependent claims 2-9, 11-15 and 17-21 are not patent eligible under the same rationale as provided for in the rejection of claims 1, 10 and 16.
Thus, claims 1-21 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.
Claim(s) 1, 3, 6, 7, 10, 12, 14-16, 18, 20, 21 are rejected under 35 U.S.C. 103 as being unpatentable over Chenchu et al., US 20210020056 A1, in view of Liu et al., ‘Dynamic Game Theoretic Neural Optimizer” (June 2021), hereinafter referred to as Chenchu and Liu, respectively.
Regarding claim 1, Chenchu discloses an anomaly detection device for an aircraft comprising (A processor-implemented system for detecting trajectory anomalies during takeoff or landing at an airdrome is disclosed – See at least ¶5):
a memory and a processor configured to (The system includes one or more processors configured by programming instructions on computer readable media – See at least ¶5)
receive satellite position data collected by the aircraft and comprising a sequence of aircraft positions defining an aircraft trajectory (The example monitoring module is configured to receive aircraft state information (e.g., position and heading, etc.) from the first aircraft at a plurality of times during performance by the first aircraft of the approved procedure – See at least ¶27. The trained models are configured to provide an expected trajectory for an aircraft at different points during performance of an approved procedure in connection with the approved runway. The historical track data from other aircraft may include position and heading data – See at least ¶49),
process the satellite position data using a plurality of different deep learning models to determine respective aircraft trajectory anomalies in a runway approach flight path for the aircraft (The example aircraft trajectory monitoring system is configured to receive a copy of an ATC clearance message (voice or text) from aircraft systems, automatically interpret the ATC clearance message to identify a specific runway to which the aircraft will land or from which the aircraft will take off, and use the identification of the specific runway to automatically identify and fetch a runway-specific model, e.g., from a trajectory models database, for use when monitoring for anomalies – See at least ¶43),
select a given deep learning model that determines the aircraft trajectory anomalies in the runway approach flight path based upon a model (The example monitoring module may select the runway-specific trained model from a trajectory models database associated with the aircraft trajectory monitoring system – See at least ¶47. Detecting, by the processor, an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level – See at least ¶78), and
generate an alert if the respective aircraft trajectory anomaly in the runway approach flight path determined by the given deep learning model exceeds a threshold (Detecting an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level; and generating, by the processor, an alert responsive to detection of the anomaly – See at least ¶84).
Chenchu fails to explicitly disclose select a given deep learning model that most accurately determines the anomalies in the runway approach flight path based upon a game theoretic model.
However, Liu teaches a game theoretic model (In this work, we address the aforementioned issues using dynamic game theory, i.e. game theoretic model, a discipline of interactive decision making built upon optimal control and game theory. Specifically, we propose to treat each layer as a player in a dynamic game connected through the network propagation. The additional dimension gained from multi-player allows us to generalize OCT-inspired methods to accept a much richer network class. Further, introducing game-theoretic analysis, e.g. information structure, provides a novel algorithmic connection between different classes of training methods from a Nash equilibria standpoint – See at least Introduction paragraph, pages 1 and 2).
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 invention of Chenchu and include the feature of a game theoretic model, as taught by Liu, yielding predictable improvements in anomaly handling and decision making and enhancing safety in aircraft operations.
Regarding claim 3, Chenchu discloses wherein the processor is further configured to generate distorted aircraft trajectory data, and process the satellite position data convolved with the distorted aircraft trajectory data using the plurality of different deep learning models to determine the aircraft trajectory anomalies (FIG. 3 is a process flow chart depicting an example process for building a runway-specific model for use in a trajectory monitoring system and an example process for using a runway-specific model to identify trajectory anomalies, i.e. distorted aircraft trajectory data – See at least ¶36. The example process further includes building the runway-specific model from the trained historical trajectory data. Building the runway-specific model may include training a machine learning model such as a support vector machine, neural network, Bayesian network, or other model. The trained model may be trained to compare a current aircraft trajectory to an expected trajectory for an aircraft at different points during performance of a specific procedure in connection with a specific runway and identify the level of deviation from the expected trajectory – See at least ¶37).
Regarding claim 6, Chenchu as modified discloses the device of claim 1, accordingly, the rejection of claim 1 above is incorporated.
Chenchu as modified does not explicitly disclose wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models.
However, Liu teaches wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models (This includes standard training methods such as SGD, RMSprop – See at least, Section 5.1, page 7, “Evaluation on Classification Datasets”).
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 invention of Chenchu and include the feature of wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Regarding claim 7, Chenchu as modified discloses the device of claim 1, accordingly, the rejection of claim 1 above is incorporated.
Chenchu as modified does not explicitly disclose wherein the processor is configured to solve the game theoretic model using a linear program.
However, Liu teaches wherein the processor is configured to solve the game theoretic model using a linear program (Computing the game-theoretic objectives requires knowing (Ft; `t;n; _n). Despite they are well-defined from 3.1, carrying Qt;n or Pt as a stage-varying function is computationally impractical even on a relatively lowdimensional system, let alone DNNs. Since the goal is to derive an incremental update given partial data at each training iteration, we can consider solving them approximately via linearization – See at least section 4.1, page 5, “Iterative Update via Linearization”).
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 invention of Chenchu and include the feature of wherein the processor is configured to solve the game theoretic model using a linear program, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Regarding claim 10, Chenchu discloses an anomaly detection method comprising (A processor-implemented system for detecting trajectory anomalies during takeoff or landing at an airdrome is disclosed – See at least ¶5):
at an anomaly detection device for an aircraft, receiving satellite position data collected by the aircraft and comprising a sequence of aircraft positions defining an aircraft trajectory (The example monitoring module is configured to receive aircraft state information (e.g., position and heading, etc.) from the first aircraft at a plurality of times during performance by the first aircraft of the approved procedure – See at least ¶27. The trained models are configured to provide an expected trajectory for an aircraft at different points during performance of an approved procedure in connection with the approved runway. The historical track data from other aircraft may include position and heading data – See at least ¶49),
processing the satellite position data using a plurality of different deep learning models to determine respective aircraft trajectory anomalies in a runway approach flight path for the aircraft (The example aircraft trajectory monitoring system is configured to receive a copy of an ATC clearance message (voice or text) from aircraft systems, automatically interpret the ATC clearance message to identify a specific runway to which the aircraft will land or from which the aircraft will take off, and use the identification of the specific runway to automatically identify and fetch a runway-specific model, e.g., from a trajectory models database, for use when monitoring for anomalies – See at least ¶43),
selecting a given deep learning model that determines the aircraft trajectory anomalies in the runway approach flight path based upon a model (The example monitoring module may select the runway-specific trained model from a trajectory models database associated with the aircraft trajectory monitoring system – See at least ¶47. Detecting, by the processor, an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level – See at least ¶78), and
generating an alert if the respective aircraft trajectory anomaly in the runway approach flight path determined by the given deep learning model exceeds a threshold (Detecting an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level; and generating, by the processor, an alert responsive to detection of the anomaly – See at least ¶84).
Chenchu fails to explicitly disclose select a given deep learning model that most accurately determines the anomalies in the runway approach flight path based upon a game theoretic model.
However, Liu teaches a game theoretic model (In this work, we address the aforementioned issues using dynamic game theory, i.e. game theoretic model, a discipline of interactive decision making built upon optimal control and game theory. Specifically, we propose to treat each layer as a player in a dynamic game connected through the network propagation. The additional dimension gained from multi-player allows us to generalize OCT-inspired methods to accept a much richer network class. Further, introducing game-theoretic analysis, e.g. information structure, provides a novel algorithmic connection between different classes of training methods from a Nash equilibria standpoint – See at least Introduction paragraph, pages 1 and 2).
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 invention of Chenchu and include the feature of a game theoretic model, as taught by Liu, yielding predictable improvements in anomaly handling and decision making and enhancing safety in aircraft operations.
Regarding claim 12, Chenchu discloses at the anomaly detection device, generating distorted aircraft trajectory data, and processing the satellite position data and distorted aircraft trajectory data using the plurality of different deep learning models to determine the aircraft trajectory anomalies (FIG. 3 is a process flow chart depicting an example process for building a runway-specific model for use in a trajectory monitoring system and an example process for using a runway-specific model to identify trajectory anomalies, i.e. distorted aircraft trajectory data – See at least ¶36. The example process further includes building the runway-specific model from the trained historical trajectory data. Building the runway-specific model may include training a machine learning model such as a support vector machine, neural network, Bayesian network, or other model. The trained model may be trained to compare a current aircraft trajectory to an expected trajectory for an aircraft at different points during performance of a specific procedure in connection with a specific runway and identify the level of deviation from the expected trajectory – See at least ¶37).
Regarding claim 14, Chenchu as modified discloses the method of claim 10, accordingly, the rejection of claim 10 above is incorporated.
Chenchu as modified does not explicitly disclose wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models.
However, Liu teaches wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models (This includes standard training methods such as SGD, RMSprop – See at least, Section 5.1, page 7, “Evaluation on Classification Datasets”).
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 invention of Chenchu and include the feature of wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Regarding claim 15, Chenchu as modified discloses the method of claim 10, accordingly, the rejection of claim 10 above is incorporated.
Chenchu as modified does not explicitly disclose wherein selecting further comprises solving the game theoretic model using a linear program.
However, Liu teaches wherein selecting further comprises solving the game theoretic model using a linear program (Computing the game-theoretic objectives requires knowing (Ft; `t;n; _n). Despite they are well-defined from 3.1, carrying Qt;n or Pt as a stage-varying function is computationally impractical even on a relatively lowdimensional system, let alone DNNs. Since the goal is to derive an incremental update given partial data at each training iteration, we can consider solving them approximately via linearization – See at least section 4.1, page 5, “Iterative Update via Linearization”).
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 invention of Chenchu and include the feature of wherein selecting further comprises solving the game theoretic model using a linear program, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Regarding claim 16, Chenchu discloses a non-transitory computer-readable medium having computer-executable instructions for causing an anomaly detection device for an aircraft to perform steps comprising (A processor-implemented system for detecting trajectory anomalies during takeoff or landing at an airdrome is disclosed – See at least ¶5):
receiving satellite position data collected by the aircraft and comprising a sequence of aircraft positions defining an aircraft trajectory (The example monitoring module is configured to receive aircraft state information (e.g., position and heading, etc.) from the first aircraft at a plurality of times during performance by the first aircraft of the approved procedure – See at least ¶27. The trained models are configured to provide an expected trajectory for an aircraft at different points during performance of an approved procedure in connection with the approved runway. The historical track data from other aircraft may include position and heading data – See at least ¶49);
processing the satellite position data using a plurality of different deep learning models to determine respective aircraft trajectory anomalies in a runway approach flight path for the aircraft (The example aircraft trajectory monitoring system is configured to receive a copy of an ATC clearance message (voice or text) from aircraft systems, automatically interpret the ATC clearance message to identify a specific runway to which the aircraft will land or from which the aircraft will take off, and use the identification of the specific runway to automatically identify and fetch a runway-specific model, e.g., from a trajectory models database, for use when monitoring for anomalies – See at least ¶43);
selecting a given deep learning model that determines the aircraft trajectory anomalies in the runway approach flight path based upon a model (The example monitoring module may select the runway-specific trained model from a trajectory models database associated with the aircraft trajectory monitoring system – See at least ¶47. Detecting, by the processor, an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level – See at least ¶78); and
generating an alert if the respective aircraft trajectory anomaly in the runway approach flight path determined by the given deep learning model exceeds a threshold (Detecting an anomaly when the trajectory of the first aircraft deviates from the expected trajectory provided by the selected model by more than a predetermined threshold level; and generating, by the processor, an alert responsive to detection of the anomaly – See at least ¶84).
Chenchu fails to explicitly disclose selecting a given deep learning model that most accurately determines the anomalies in the runway approach flight path based upon a game theoretic model.
However, Liu teaches a game theoretic model (In this work, we address the aforementioned issues using dynamic game theory, i.e. game theoretic model, a discipline of interactive decision making built upon optimal control and game theory. Specifically, we propose to treat each layer as a player in a dynamic game connected through the network propagation. The additional dimension gained from multi-player allows us to generalize OCT-inspired methods to accept a much richer network class. Further, introducing game-theoretic analysis, e.g. information structure, provides a novel algorithmic connection between different classes of training methods from a Nash equilibria standpoint – See at least Introduction paragraph, pages 1 and 2).
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 invention of Chenchu and include the feature of a game theoretic model, as taught by Liu, yielding predictable improvements in anomaly handling and decision making and enhancing safety in aircraft operations.
Regarding claim 18, Chenchu discloses a causing the anomaly detection device to perform steps comprising generating distorted aircraft trajectory data, and processing the satellite position data and distorted aircraft trajectory data using the plurality of different deep learning models to determine the aircraft trajectory anomalies (FIG. 3 is a process flow chart depicting an example process for building a runway-specific model for use in a trajectory monitoring system and an example process for using a runway-specific model to identify trajectory anomalies, i.e. distorted aircraft trajectory data – See at least ¶36. The example process further includes building the runway-specific model from the trained historical trajectory data. Building the runway-specific model may include training a machine learning model such as a support vector machine, neural network, Bayesian network, or other model. The trained model may be trained to compare a current aircraft trajectory to an expected trajectory for an aircraft at different points during performance of a specific procedure in connection with a specific runway and identify the level of deviation from the expected trajectory – See at least ¶37).
Regarding claim 20, Chenchu as modified discloses the medium of claim 16, accordingly, the rejection of claim 16 above is incorporated.
Chenchu as modified does not explicitly disclose wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models.
However, Liu teaches wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models (This includes standard training methods such as SGD, RMSprop – See at least, Section 5.1, page 7, “Evaluation on Classification Datasets”).
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 invention of Chenchu and include the feature of wherein the plurality of different deep learning models comprises at least one of Adaptive Moment Estimation (ADAM), Stochastic Gradient Descent with Momentum (SGDM), and RMSProp deep learning models, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Regarding claim 21, Chenchu as modified discloses the medium of claim 16, accordingly, the rejection of claim 16 above is incorporated.
Chenchu as modified does not explicitly disclose wherein selecting further comprises solving the game theoretic model using a linear program.
However, Liu teaches wherein selecting further comprises solving the game theoretic model using a linear program (Computing the game-theoretic objectives requires knowing (Ft; `t;n; _n). Despite they are well-defined from 3.1, carrying Qt;n or Pt as a stage-varying function is computationally impractical even on a relatively lowdimensional system, let alone DNNs. Since the goal is to derive an incremental update given partial data at each training iteration, we can consider solving them approximately via linearization – See at least section 4.1, page 5, “Iterative Update via Linearization”).
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 invention of Chenchu and include the feature of wherein selecting further comprises solving the game theoretic model using a linear program, as taught by Liu, to enhance the robustness or convergence of the optimization process (See Liu, Section 5.2, “Game-Theoretic Applications”).
Claims 2, 11 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Chenchu et al., US 20210020056 A1, in view of Liu et al., ‘Dynamic Game Theoretic Neural Optimizer” (June 2021) as applied to claims 1, 10 and 16 above and further in view of Nanduri et al., “Anomaly detection in aircraft data using Recurrent Neural Networks (RNN)” In 2016 Integrated Communications Navigation and Surveillance (ICNS): April 2016; pgs. 8, hereinafter referred to as Chenchu, Liu and Nanduri, respectively.
Regarding claim 2, the combination of Chenchu and Liu fails to explicitly disclose wherein the aircraft trajectory comprises a state-vector; and wherein the processor is further configured to resample the state-vector to normalize timing and velocity.
However, Nanduri teaches wherein the aircraft trajectory comprises a state-vector; and wherein the processor is further configured to resample the state-vector to normalize timing and velocity (Flight parameters are checked for exceedance of predefined limits under certain conditions. The limits and conditions are defined in advance by domain experts based on airlines standard operating procedures. Example events include: pitch and speed during takeoff – See at least page 2, ¶ “Exceedance Detection” paragraph. Clustering based anomaly detection initially converts the raw data into time series data. In order to map data into comparable vectors, i.e. state vectors, in the high dimensional space, these time series data from different flights are anchored by a specific event to make temporal patterns comparable. Then every flight parameter is sampled, i.e. resample, at fixed intervals by time, distance or other reference from the reference event. All sampled values are arranged to form a high dimensional vector for each flight – See at least page 2, ¶ “Clustering Based Anomaly Detection (ClusterAD).
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 combination of Chenchu and Liu and include the feature of wherein the aircraft trajectory comprises a state-vector; and wherein the processor is further configured to resample the state-vector to normalize timing and velocity, as taught by Nanduri, to provide real-time anomaly detection on a flight deck (See Nanduri, Page 1, “Abstract”).
Regarding claim 11, the combination of Chenchu and Liu fails to explicitly disclose wherein the aircraft trajectory comprises a vector; and further comprising, at the anomaly detection device, resampling the vector to normalize timing and velocity.
However, Nanduri teaches wherein the aircraft trajectory comprises a vector; and further comprising, at the anomaly detection device, resampling the vector to normalize timing and velocity (Flight parameters are checked for exceedance of predefined limits under certain conditions. The limits and conditions are defined in advance by domain experts based on airlines standard operating procedures. Example events include: pitch and speed during takeoff – See at least page 2, ¶ “Exceedance Detection” paragraph. Clustering based anomaly detection initially converts the raw data into time series data. In order to map data into comparable vectors, i.e. state vectors, in the high dimensional space, these time series data from different flights are anchored by a specific event to make temporal patterns comparable. Then every flight parameter is sampled, i.e. resample, at fixed intervals by time, distance or other reference from the reference event. All sampled values are arranged to form a high dimensional vector for each flight – See at least page 2, ¶ “Clustering Based Anomaly Detection (ClusterAD).
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 combination of Chenchu and Liu and include the feature of wherein the aircraft trajectory comprises a vector; and further comprising, at the anomaly detection device, resampling the vector to normalize timing and velocity, as taught by Nanduri, to provide real-time anomaly detection on a flight deck (See Nanduri, Page 1, “Abstract”).
Regarding claim 17, the combination of Chenchu and Liu fails to explicitly disclose wherein the aircraft trajectory comprises a vector; and further having computer-executable instructions for causing the anomaly detection device to perform a step of resampling the vector to normalize timing and velocity.
However, Nanduri teaches wherein the aircraft trajectory comprises a vector; and further having computer-executable instructions for causing the anomaly detection device to perform a step of resampling the vector to normalize timing and velocity (Flight parameters are checked for exceedance of predefined limits under certain conditions. The limits and conditions are defined in advance by domain experts based on airlines standard operating procedures. Example events include: pitch and speed during takeoff – See at least page 2, ¶ “Exceedance Detection” paragraph. Clustering based anomaly detection initially converts the raw data into time series data. In order to map data into comparable vectors, i.e. state vectors, in the high dimensional space, these time series data from different flights are anchored by a specific event to make temporal patterns comparable. Then every flight parameter is sampled, i.e. resample, at fixed intervals by time, distance or other reference from the reference event. All sampled values are arranged to form a high dimensional vector for each flight – See at least page 2, ¶ “Clustering Based Anomaly Detection (ClusterAD).
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 combination of Chenchu and Liu and include the feature of wherein the aircraft trajectory comprises a vector; and further having computer-executable instructions for causing the anomaly detection device to perform a step of resampling the vector to normalize timing and velocity, as taught by Nanduri, to provide real-time anomaly detection on a flight deck (See Nanduri, Page 1, “Abstract”).
Claims 4, 5, 13 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Chenchu et al., US 20210020056 A1, in view of Liu et al., ‘Dynamic Game Theoretic Neural Optimizer” (June 2021), as applied to claims 1, 10 and 16 above and further in view of Zhang et al., WO2021225741 , hereinafter referred to as Chenchu, Liu and Zhang, respectively.
Regarding claim 4, the combination of Chenchu and Liu fails to explicitly disclose wherein the processor is further configured to implement a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model.
However, Zhang teaches wherein the processor is further configured to implement a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model (In a VAE, the performance of any imputation or prediction performed based on the latent vector depends on the dimensionality of the latent space – See at least ¶13. For instance, a random or unobserved value of the latent vectorH can be input to the second decoder in order to generate a new instance of the feature vector X that was not observed in the training data. E.g. this could be used to generate a fictional face for use in a move or game, or to generate details of a functional patient for training or study purposes – See at least ¶71).
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 combination of Chenchu and Liu and include the feature of wherein the processor is further configured to implement a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model, as taught by Zhang, to provide a machine learning model which can handle mixed types of data with reduced computational complexity for a given performance (See Zhang ¶14).
Regarding claim 5, the combination of Chenchu and Lui fails to explicitly disclose wherein the VAE comprises an encoder configured to generate a mean vector and a standard deviation vector from the satellite position data, and generate the latent vector from the mean vector and the standard deviation vector.
However, Zhang teaches wherein the VAE comprises an encoder configured to generate a mean vector and a standard deviation vector from the satellite position data, and generate the latent vector from the mean vector and the standard deviation vector (In embodiments each of the latent representations is one-dimensional, i.e. consists of only a single latent variable (element). Note however this does not imply the latent variable is a modelled only as simple, fixed scalar value. Rather, as the auto encoder is a variation auto-encoder, then for each latent variable the encoder learns a statistical or probabilistic distribution, and the value input to the decoder is a random sample from the distribution. This means that for each individual element of latent space, the encoder learns one or more parameters of the respective distribution, e.g. a measure of centre point and spread of the distribution. For instance each latent variable (a single dimension) may be modelled in the encoder by a respective mean value standard deviation ad – See at least ¶63).
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 combination of Chenchu and Liu and include the feature of wherein the VAE comprises an encoder configured to generate a mean vector and a standard deviation vector from the satellite position data, and generate the latent vector from the mean vector and the standard deviation vector, as taught by Zhang, to provide a machine learning model which can handle mixed types of data with reduced computational complexity for a given performance (See Zhang ¶14).
Regarding claim 13, the combination of Chenchu and Liu fails to explicitly disclose at the anomaly detection device, implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model.
However, Zhang teaches at the anomaly detection device, implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model (In a VAE, the performance of any imputation or prediction performed based on the latent vector depends on the dimensionality of the latent space – See at least ¶13. For instance, a random or unobserved value of the latent vectorH can be input to the second decoder in order to generate a new instance of the feature vector X that was not observed in the training data. E.g. this could be used to generate a fictional face for use in a move or game, or to generate details of a functional patient for training or study purposes – See at least ¶71).
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 combination of Chenchu and Liu and include the feature of at the anomaly detection device, implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model, as taught by Zhang, to provide a machine learning model which can handle mixed types of data with reduced computational complexity for a given performance (See Zhang ¶14).
Regarding claim 19, the combination of Chenchu and Liu fails to explicitly disclose computer-executable instructions for causing the anomaly detection device to perform a step of implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model.
However, Zhang teaches computer-executable instructions for causing the anomaly detection device to perform a step of implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model (In a VAE, the performance of any imputation or prediction performed based on the latent vector depends on the dimensionality of the latent space – See at least ¶13. For instance, a random or unobserved value of the latent vectorH can be input to the second decoder in order to generate a new instance of the feature vector X that was not observed in the training data. E.g. this could be used to generate a fictional face for use in a move or game, or to generate details of a functional patient for training or study purposes – See at least ¶71).
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 combination of Chenchu and Liu and include the feature of computer-executable instructions for causing the anomaly detection device to perform a step of implementing a variational autoencoder (VAE) to analyze a latent space vector to gather confidence metrics for the game theoretic model, as taught by Zhang, to provide a machine learning model which can handle mixed types of data with reduced computational complexity for a given performance (See Zhang ¶14).
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Chenchu et al., US 20210020056 A1, in view of Liu et al., ‘Dynamic Game Theoretic Neural Optimizer” (June 2021), as applied to claim 1 above and further in view of Caule et al., US 20130211632 A1, hereinafter referred to as Chenchu, Liu and Caule, respectively.
Regarding claim 8, the combination of Chenchu and Liu fails to explicitly disclose wherein the satellite position data comprises Global Positioning System (GPS) data.
However, Caule teaches wherein the satellite position data comprises Global Positioning System (GPS) data (In the second type of algorithm, the triggering of the warning is based on information contained in one or more databases (for example airport runway) and geographic positioning equipment on the aircraft (for example GPS, or Global Positioning System, equipment – See at least ¶9).
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 combination of Chenchu and Liu and include the feature of wherein the satellite position data comprises Global Positioning System (GPS) data, as taught by Caule, to generate warnings (audible and/or visual) intended for the crew of the aircraft, on the basis of collision warning algorithms that can be “reactive” or “predictive” (See Caule ¶8).
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Chenchu et al., US 20210020056 A1, in view of Liu et al., ‘Dynamic Game Theoretic Neural Optimizer” (June 2021), as applied to claim 1 above and further in view of Hardman et al., US 20080215192 A1, hereinafter referred to as Chenchu, Liu and Hardman, respectively.
Regarding claim 9, the combination of Chenchu and Liu fails to explicitly disclose a housing carrying the memory and processor, the housing configured to be mounted within the aircraft.
However, Hardman teaches a housing carrying the memory and processor, the housing configured to be mounted within the aircraft (a housing that retains the processor, the memory, the housing physically configured to replace a legacy device mounted in a cockpit instrument panel of the aircraft – See at least claim 1 of Hardman and FIG 1).
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 combination of Chenchu and Liu and include the feature of a housing carrying the memory and processor, the housing configured to be mounted within the aircraft, as taught by Hardman, to enhance the functionality of the instrument panel (See Hardman ¶1).
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MAHMOUD M KAZIMI whose telephone number is (571)272-3436. The examiner can normally be reached M-F 7am-5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Erin Bishop can be reached at 5712703713. 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.
/M.M.K./Examiner, Art Unit 3665
/DONALD J WALLACE/Primary Examiner, Art Unit 3665