# US 2021/0082292 A1 · Unsupervised anomaly detection for autonomous vehicles

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0082292 A1
- Document ID: US20210082292A1
- Status: Published application
- Kind: A1
- Application number: 16/887,194
- Filed: 2020-05-29
- Published: 2021-03-18
- Earliest claimed: 2019-09-13
- Assignees: Wing Aviation LLC
- Applicants: Wing Aviation LLC
- Inventors: Vikas Sindhwani; Hakim Sidahmed; Krzysztof Choromanski; Brandon L. Jones
- Web page: https://patentgrep.com/patent/US20210082292A1

## CPC

- G08G 5/55: for a single aircraft
- B64F 5/60: Testing or inspecting aircraft components or systems
- B64U 10/25: Fixed-wing aircraft (VTOL aircraft B64U10/20)
- B64U 50/13: using external fans or propellers
- G05D 1/0011: {associated with a remote control arrangement}
- G05D 1/0055: {with safety arrangements}
- G05D 1/105: {specially adapted for unpowered flight, e.g. glider, parachuting, forced landing (parachutes per se B64D17/00)}
- G06F 11/0736: {in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function (testing or monitoring of automated control systems G05B23/02)}
- G06F 11/0751: {Error or fault detection not based on redundancy (power supply failures G06F1/30; network fault management H04L41/06)}
- G06F 11/0772: {Means for error signaling, e.g. using interrupts, exception flags, dedicated error registers}
- G06F 11/079: {Root cause analysis, i.e. error or fault diagnosis (in a hardware test environment G06F11/22; in a software test environment G06F11/36)}
- G06F 11/3013: {where the computing system is an embedded system, i.e. a combination of hardware and software dedicated to perform a certain function in mobile devices, printers, automotive or aircraft systems (testing or monitoring of control systems or parts thereof G05B23/02)}
- G06F 11/3447: {Performance evaluation by modeling}
- G06N 20/00: Machine learning
- G08G 5/21: located onboard the aircraft
- G08G 5/53: for cruising
- G08G 5/57: for unmanned aircraft
- G08G 5/58: for emergency situations, e.g. hijacking or bird strikes
- B64U 2201/20: Remote controls
- G06F 2201/81: Threshold
- G07C 5/0808: {Diagnosing performance data (testing of vehicles G01M17/00; testing of electrical installation on vehicles G01R31/005)}
- G08G: TRAFFIC CONTROL SYSTEMS (guiding railway traffic, ensuring the safety of railway traffic B61L; arrangement of road signs or traffic signals E01F9/00; radar or analogous systems, sonar systems, lidar systems specially adapted for traffic control G01S13/91, G01S15/88, G01S17/88; {radar or analogous systems, sonar systems, lidar systems specially adapted for anti-collision purposes G01S13/93, G01S15/93, G01S17/93})
- B64F: GROUND OR AIRCRAFT-CARRIER-DECK INSTALLATIONS SPECIALLY ADAPTED FOR USE IN CONNECTION WITH AIRCRAFT; DESIGNING, MANUFACTURING, ASSEMBLING, CLEANING, MAINTAINING OR REPAIRING AIRCRAFT, NOT OTHERWISE PROVIDED FOR; HANDLING, TRANSPORTING, TESTING OR INSPECTING AIRCRAFT COMPONENTS, NOT OTHERWISE PROVIDED FOR
- B64U: UNMANNED AERIAL VEHICLES [UAV]; EQUIPMENT THEREFOR
- G05D: SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G07C: TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE

## Abstract

In some embodiments, techniques are provided for analyzing time series data to detect anomalies. In some embodiments, the time series data is processed using a machine learning model. In some embodiments, the machine learning model is trained in an unsupervised manner on large amounts of previous time series data, thus allowing highly accurate models to be created from novel data. In some embodiments, training of the machine learning model alternates between a fitting optimization and a trimming optimization to allow large amounts of training data that includes untagged anomalous records to be processed. Because a machine learning model is used, anomalies can be detected within complex systems, including but not limited to autonomous vehicles such as unmanned aerial vehicles. When anomalies are detected, commands can be transmitted to the monitored system (such as an autonomous vehicle) to respond to the anomaly.

## Claims

### Claim 1

A non-transitory computer-readable medium having logic stored thereon that, in response to execution by one or more processors of a computing system, cause the computing system to perform actions for detecting anomalies in time series data records, the actions comprising:
receiving, by the computing system, a plurality of time series data records;
initializing, by the computing system, a machine learning model having a set of fitting weights and a set of time series data record weights;
optimizing, by the computing system, the set of fitting weights of the machine learning model while holding the set of trip weights of the machine learning model constant;
optimizing, by the computing system, the set of time series data record weights while holding the set of fitting weights constant; and
storing, by the computing system, the optimized time series data record weights and the optimized fitting weights for use in monitoring for anomalies.

### Claim 2 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the actions further comprise determining an anomaly threshold value.

### Claim 3 (depends on 1)

The non-transitory computer-readable medium of claim 1, wherein the plurality of time series data records include telemetry information from one or more autonomous vehicles.

### Claim 4 (depends on 3)

The non-transitory computer-readable medium of claim 3, wherein the one or more autonomous vehicles includes at least one unmanned aerial vehicle (UAV).

### Claim 5 (depends on 4)

The non-transitory computer-readable medium of claim 4, wherein the UAV includes at least one horizontal propulsion unit and at least one vertical propulsion unit.

### Claim 6 (depends on 4)

The non-transitory computer-readable medium of claim 4, wherein the telemetry information includes information from at least two of a takeoff portion of a trip, a cruise portion of the trip, and a landing portion of the trip.

### Claim 7 (depends on 4)

The non-transitory computer-readable medium of claim 4, wherein the time series data records include data that represents at least one of a motor speed, a control surface servo position, a linear velocity, an orientation, and an angular velocity.

### Claim 8 (depends on 3)

The non-transitory computer-readable medium of claim 3, wherein the actions further comprise:
receiving a new time series data record from an autonomous vehicle; and
processing the new time series data record with the machine learning model to detect anomalies.

### Claim 9 (depends on 8)

The non-transitory computer-readable medium of claim 8, wherein the actions further comprise, in response to detecting an anomaly:
determining an action to be taken to address the anomaly; and
transmitting a command to the autonomous vehicle to cause the autonomous vehicle to perform the action to address the anomaly.

### Claim 10 (depends on 9)

The non-transitory computer-readable medium of claim 9, wherein the action to address the anomaly is at least one of rescheduling a future trip, navigating to an emergency repair location, immediately performing a landing procedure, and accepting remote control by a human operator.

### Claim 11

A method for detecting anomalies in time series data records, the method comprising:
receiving, by a computing device, a time series data record from a monitored system;
processing, by the computing device, the time series data record using a machine learning model to generate an anomaly score, wherein the machine learning model was trained on a plurality of previous time series data records;
comparing, by the computing device, the anomaly score to an anomaly threshold value; and
in response to determining that the anomaly score is greater than the anomaly threshold value:
determining, by the computing device, an action to be taken to address the anomaly; and
transmitting, by the computing device, a command to the monitored system to cause the monitored system to perform the action to address the anomaly.

### Claim 12 (depends on 11)

The method of claim 11, wherein the monitored system is an autonomous vehicle.

### Claim 13 (depends on 12)

The method of claim 12, wherein the autonomous vehicle is an unmanned aerial vehicle (UAV).

### Claim 14 (depends on 13)

The method of claim 13, wherein the UAV includes at least one horizontal propulsion unit and at least one vertical propulsion unit.

### Claim 15 (depends on 13)

The method of claim 13, wherein the time series data record includes telemetry information from at least two of a takeoff portion of a trip, a cruise portion of the trip, and a landing portion of the trip.

### Claim 16 (depends on 13)

The method of claim 13, wherein the time series data record includes data representing at least one of a motor speed, a control surface servo position, a linear velocity, an orientation, and an angular velocity.

### Claim 17 (depends on 13)

The method of claim 13, wherein the action is at least one of rescheduling a future trip, navigating to an emergency repair location, and immediately performing a landing procedure.

### Claim 18 (depends on 11)

The method of claim 11, further comprising training the machine learning model by performing actions including:
collecting a plurality of time series data records;
initializing the machine learning model to have a set of fitting weights and a set of time series data record weights;
optimizing the set of fitting weights while holding the set of time series data record weights constant; and
optimizing the set of time series data record weights while holding the set of fitting weights constant.

### Claim 19 (depends on 18)

The method of claim 18, wherein training the machine learning model further includes determining the anomaly threshold value using the plurality of time series data records.

### Claim 20

A system, comprising at least one computing device that includes:
at least one processor; and
a non-transitory computer-readable medium having logic stored thereon that, in response to execution by the at least one processor, causes the system to perform actions comprising:
receiving a time series data record from a monitored system;
processing the time series data record using a machine learning model to generate an anomaly score, wherein the machine learning model was trained on a plurality of previous time series data records;
comparing the anomaly score to an anomaly threshold value; and
in response to determining that the anomaly score is greater than the anomaly threshold value:
determining an action to be taken to address the anomaly; and
transmitting a command to the monitored system to cause the monitored system to perform the action to address the anomaly.
