Method of parallelized occupancy grid collision checking

A computer-implemented method comprising:

• identifying, based upon perception sensor data of a plurality of perception sensors, a plurality of actors in an environment of an autonomous vehicle;
• generating a respective anticipated trajectory on a canvas for each actor of the plurality of actors;
• generating a plurality of anticipated trajectories of the autonomous vehicle, wherein each anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle is generated in a separate canvas;
• performing a binary AND operation between each pixel of the canvas having the respective anticipated trajectory for each actor of the plurality of actors and each pixel of a canvas having an anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle;
• based upon the performed binary AND operation, identifying one or more pixels in which the anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle having collision with at least one actor of the plurality of actors;
• and
• discarding the anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle.

An autonomous vehicle comprising:

• a plurality of perception sensors;
• at least one memory configured to store machine executable instructions;
• and
• at least one processor coupled to the at least one memory and configured to execute the machine executable instructions to configure the at least one processor to:
• identify, based upon perception sensor data of the plurality of perception sensors, a plurality of actors in an environment of the autonomous vehicle;
• generate a respective anticipated trajectory on a canvas for each actor of the plurality of actors;
• generate a plurality of anticipated trajectories of the autonomous vehicle, wherein each anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle is generated in a separate canvas;
• perform a binary AND operation between each pixel of the canvas having the respective anticipated trajectory for each actor of the plurality of actors and each pixel of a canvas having an anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle;
• based upon the performed binary AND operation, identify one or more pixels in which the anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle having collision with at least one actor of the plurality of actors;
• and
• discard the anticipated trajectory of the plurality of anticipated trajectories of the autonomous vehicle.