US 12,228,419 B2
Systematic fault detection in vehicle control systems
Guillermo Duenas Arana, San Francisco, CA (US); Soroush Dean Khadem, Boulder, CO (US); and Till Kroeger, San Francisco, CA (US)
Assigned to Zoox, Inc., Foster City, CA (US)
Filed by Zoox, Inc., Foster City, CA (US)
Filed on Jul. 29, 2021, as Appl. No. 17/388,974.
Prior Publication US 2023/0035780 A1, Feb. 2, 2023
Int. Cl. G01C 21/00 (2006.01); G01C 21/36 (2006.01); G05D 1/00 (2006.01)
CPC G01C 21/3811 (2020.08) [G01C 21/367 (2013.01); G01C 21/3833 (2020.08); G01C 21/387 (2020.08); G05D 1/0221 (2013.01); G05D 1/0276 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computer-executable instructions that, when executed, cause the system to perform operations comprising:
receiving log data associated with an autonomous vehicle operating in a physical environment;
receiving map data associated with the physical environment;
injecting, into the map data, a systematic fault, wherein the systematic fault comprises one or more of a simulated addition, a simulated deletion, a simulated rotation, or a simulated shift associated with a portion of the physical environment corresponding to a lane marking, and
further wherein, when instantiated in a simulated environment representing at least the portion of the physical environment corresponding to the lane marking, the systematic fault affects a first semantic class of a semantic-based localization system associated with a simulated vehicle controller and leaves unaffected a second semantic class of the semantic-based localization system associated with the simulated vehicle controller;
generating, based at least in part on the log data and the systematic fault, the simulated environment representing the at least the portion of the physical environment corresponding to the lane marking;
determining a simulated trajectory for the simulated vehicle controller to cause a simulated vehicle to traverse the simulated environment;
evaluating, as a simulation, simulated motion of the simulated vehicle following the simulated trajectory;
determining, based on the simulation and at a simulated time at which the simulated vehicle controller causes the simulated vehicle to traverse the systematic fault, a simulated position of the simulated vehicle in the simulated environment;
determining, based on the log data and at a time corresponding to the simulated time, a physical position of the autonomous vehicle in the physical environment;
determining a difference between the physical position of the autonomous vehicle in the physical environment and the simulated position of the simulated vehicle in the simulated environment; and
validating the simulated vehicle controller based at least in part on the difference.