US 12,072,419 B2
Procedurally generated three-dimensional environment for use in autonomous vehicle simulations
Benjamin Goldstein, San Francisco, CA (US); Gen Nishida, Daly City, CA (US); and Eileen Bai, San Francisco, CA (US)
Assigned to GM Cruise Holdings LLC, San Francisco, CA (US)
Filed by GM Cruise Holdings LLC, San Francisco, CA (US)
Filed on Dec. 17, 2020, as Appl. No. 17/125,558.
Prior Publication US 2022/0196839 A1, Jun. 23, 2022
Int. Cl. G01S 17/00 (2020.01); G01S 17/89 (2020.01); G06T 17/20 (2006.01)
CPC G01S 17/89 (2013.01) [G06T 17/20 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method for generating a synthetic map, comprising:
collecting, by an autonomous vehicle traversing a real-world environment, sensor data corresponding with a three-dimensional (3D) space in the real-world scene and first outputs of an autonomous vehicle stack as the autonomous vehicle traverses the real-world environment;
generating a 3D mesh based on the sensor data;
generating one or more synthetic 3D objects based on the 3D mesh and the sensor data;
generating a 3D synthetic environment comprising the one or more synthetic 3D objects, wherein the 3D synthetic environment is generated based on the 3D mesh;
simulating navigation of a simulated autonomous vehicle, which is operating using the autonomous vehicle stack, through at least a portion of the 3D map to generate a second outputs of the autonomous vehicle stack; and
comparing the first outputs of the autonomous vehicle stack to the second outputs of the autonomous vehicle stack to determine a validity of the 3D synthetic environment.