Systems and methods for injecting faults into an autonomy system

1. A computer-implemented method comprising:

obtaining data for one or more data variables related to autonomous operation of a vehicle in a test environment being facilitated by an autonomy system of the vehicle;

based on the data, evaluating of a set of predefined fault rules, wherein each of the predefined fault rules comprises (i) a set of predefined criteria related to the one or more data variables and (ii) a predefined fault that is to be injected into the autonomy system of the vehicle when the set of predefined criteria are determined to be satisfied, and wherein the predefined fault includes at least one of: a planning fault and a control fault,

wherein the set of predefined criteria comprises at least one (i) temporal criterion, (ii) ambient-condition criterion, and (iii) operator criterion, and wherein the predefined criteria are satisfied when a) an operator of the vehicle is associated with at least one of a required experience level and a fault history for the operator criterion, b) the vehicle is on a test mission at a pre-determined time for the temporal criterion, and c) at least one of a weather condition, a lighting condition, and a presence of an object is present in surroundings of the vehicle for the ambient-condition criterion;

in response to evaluating the set of predefined fault rules, injecting the predefined fault into the autonomy system of the vehicle; and

capturing data indicative of a response by a response mechanism of the vehicle to the vehicle autonomously operating in accordance with the predefined fault.

2. The computer-implemented method of claim 1 , wherein obtaining data for the one or more data variables related to autonomous operation of the vehicle comprises obtaining, from one or more sensors of the vehicle while the vehicle is operating in the test environment, sensor data related to the autonomous operation of the vehicle within the test environment.

3. The computer-implemented method of claim 1 , wherein the autonomy system comprises one or more of (i) a sensor subsystem, (ii) a perception subsystem, (iii) a prediction subsystem, (iv) a planning subsystem, or (v) a control subsystem.

4. The computer-implemented method of claim 3 , wherein the autonomy system further comprises a vehicle interface subsystem, and wherein injecting the predefined fault into the autonomy system of the vehicle comprises injecting at least one of (i) the planning fault into data that is input into the control subsystem, wherein the planning fault is associated with a planned trajectory of the vehicle and (ii) the control fault into data that is input into the vehicle interface subsystem, wherein the control fault is associated with an actuation of at least one physical component of the vehicle.

5. The computer-implemented method of claim 1 , wherein the autonomy system uses sensor data to generate derived data that facilitates autonomous operation of the vehicle in the test environment, wherein the derived data is generated by the autonomy system according to the sensor data, and wherein injecting the predefined fault into the autonomy system of the vehicle comprises:

modifying the derived data such that at least one aspect of the derived data differs from the derived data that was originally generated by the autonomy system of the vehicle; and

causing the vehicle to autonomously operate in accordance with the derived data.

6. The computer-implemented method of claim 5 , wherein modifying the derived data such that at least one aspect of the derived data differs from the derived data that was originally generated by the autonomy system of the vehicle comprises:

modifying at least one of (i) derived data generated by the autonomy system of the vehicle that is indicative of a perception of the vehicle of the test environment, (ii) derived data generated by the autonomy system of the vehicle that is indicative of a predicted future behavior of one or more objects perceived by the vehicle in the test environment, (iii) derived data generated by the autonomy system of the vehicle that defines a plan for the autonomous operation of the vehicle in the future, or (iv) derived data generated by the autonomy system of the vehicle that defines instructions for executing the plan.

7. The computer-implemented method of claim 1 , wherein capturing data indicative of a response by a response mechanism of the vehicle comprises capturing data indicative of a response by a computer-based response mechanism of the vehicle to the vehicle autonomously operating in accordance with the at least one injected fault.

8. A non-transitory computer-readable medium comprising program instructions stored thereon that are executable by at least one processor such that a computing system is configured to:

obtain data for one or more data variables related to autonomous operation of a vehicle in a test environment being facilitated by an autonomy system of the vehicle;

based on the data, evaluate a set of predefined fault rules, wherein each of the predefined fault rules comprises (i) a set of predefined criteria related to the one or more data variables and (ii) a predefined fault that is to be injected into the autonomy system of the vehicle when the set of predefined criteria are determined to be satisfied,

wherein the predefined fault includes at least one of: a planning fault and a control fault,

wherein the set of predefined criteria comprises at least one (i) temporal criterion, (ii) ambient-condition criterion, and (iii) operator criterion, and wherein the predefined criteria are satisfied when a) an operator of the vehicle is associated with at least one of a required experience level and a fault history for the operator criterion, b) the vehicle is on a test mission at a pre-determined time for the temporal criterion, and c) at least one of a weather condition, a lighting condition, and a presence of an object is present in surroundings of the vehicle for the ambient-condition criterion;

responsive to evaluating the set of predefined fault rules, inject the predefined fault into the autonomy system of the vehicle; and

capture data indicative of a response by a response mechanism of the vehicle to the vehicle autonomously operating in accordance with the predefined fault.

9. The non-transitory computer-readable medium of claim 8 , wherein the program instructions comprise program instructions to:

obtain, from one or more sensors of the vehicle while the vehicle is operating in the test environment, sensor data related to autonomous operation of the vehicle within the test environment.

10. The non-transitory computer-readable medium of claim 8 , wherein the autonomy system comprises one or more of (i) a sensor subsystem, (ii) a perception subsystem, (iii) a prediction subsystem, (iv) a planning subsystem, or (v) a control subsystem.

11. The non-transitory computer-readable medium of claim 10 , wherein the autonomy system further comprises a vehicle interface subsystem, and wherein the program instructions comprise program instructions to:

inject at least one of (i) the planning fault into data that is input into the control subsystem, wherein the planning fault is associated with a planned trajectory of the vehicle, and (ii) the control fault into data that is input into the vehicle interface subsystem, wherein the control fault is associated with an actuation of at least one physical component of the vehicle.

12. The non-transitory computer-readable medium of claim 8 , wherein the autonomy system uses sensor data to generate derived data that facilitates autonomous operation of the vehicle in the test environment, wherein the derived data is generated by the autonomy system according to the sensor data, and wherein the program instructions comprise program instructions to:

modify the derived data such that at least one aspect of the derived data differs from the derived data that was originally generated by the autonomy system of the vehicle; and

cause the vehicle to autonomously operate in accordance with the derived data.

13. The non-transitory computer-readable medium of claim 12 , wherein the program instructions comprise program instructions to:

modify at least one of (i) the derived data generated by the autonomy system of the vehicle that is indicative of a perception of the vehicle of the test environment, (ii) the derived data generated by the autonomy system of the vehicle that is indicative of a predicted future behavior of one or more objects perceived by the vehicle in the test environment, (iii) the derived data generated by the autonomy system of the vehicle that defines a plan for the autonomous operation of the vehicle in the future, or (iv) the derived data generated by the autonomy system of the vehicle that defines instructions for executing the plan.

14. The computer-readable medium of claim 8 , wherein the program instructions comprise program instructions to:

capture data indicative of a response by a computer-based response mechanism of the vehicle to the vehicle autonomously operating in accordance with the at least one injected fault.

15. A computing system comprising:

at least one processor;

a non-transitory computer-readable medium; and

program instructions stored on the non-transitory computer-readable medium that are executable by the at least one processor such that the computing system is configured to:

obtain data for one or more data variables related to autonomous operation of a vehicle in a test environment being facilitated by an autonomy system of the vehicle;

based on the data, evaluate of a set of predefined fault rules, wherein each of the predefined fault rules comprises (i) a set of predefined criteria related to the one or more data variables and (ii) a predefined fault that is to be injected into the autonomy system of the vehicle when the set of predefined criteria are determined to be satisfied,

wherein the predefined fault includes at least one of: a planning fault and a control fault,

wherein the set of predefined criteria comprises at least one a) temporal criterion, (ii) ambient-condition criterion, and (iii) operator criterion, and wherein the predefined criteria are satisfied when a) an operator of the vehicle is associated with at least one of a required experience level and a fault history for the operator criterion, b) the vehicle is on a test mission at a pre-determined time for the temporal criterion, and c) at least one of a weather condition, a lighting condition, and a presence of an object is present in surroundings of the vehicle for the ambient-condition criterion;

responsive to evaluating the set of predefined fault rules, inject the predefined fault into the autonomy system of the vehicle; and

capture data indicative of a response by a response mechanism of the vehicle to the vehicle autonomously operating in accordance with the predefined fault.

16. The computing system of claim 15 , wherein the autonomy system of the vehicle further comprises at least one of a control subsystem and a vehicle interface subsystem, and wherein derived data is generated by the autonomy system according to sensor data, and wherein the program instructions comprise program instructions to:

inject at least one of: (i) the planning fault into data that is input into a control subsystem, wherein the planning fault is associated with a planned trajectory of the vehicle and (ii) the control fault into data that is input into the vehicle interface subsystem, wherein the control fault is associated with an actuation of at least one physical component of the vehicle;

modify the derived data such that at least one aspect of the derived data differs from the derived data that was originally generated by the autonomy system of the vehicle; and

cause the vehicle to autonomously operate in accordance with the derived data.

17. The computing system of claim 16 , wherein the program instructions comprise program instructions to:

modify at least one of (i) derived data generated by the autonomy system of the vehicle that is indicative of a perception of the vehicle of the test environment, (ii) derived data generated by the autonomy system of the vehicle that is indicative of a predicted future behavior of one or more objects perceived by the vehicle in the test environment, (iii) derived data generated by the autonomy system of the vehicle that defines a plan for autonomous operation of the vehicle in the future, or (iv) derived data generated by the autonomy system of the vehicle that defines instructions for executing the plan.

18. The computing system of claim 15 , wherein the program instructions that are executable by the at least one processor such that the computing system is configured to capture data indicative of a response by a response mechanism of the vehicle comprise program instructions that are executable by the at least one processor such that the computing system is configured to:

capture data indicative of a response by a computer-based response mechanism of the vehicle to the vehicle autonomously operating in accordance with the at least one injected fault.