CPC B60W 30/09 (2013.01) [B60W 30/0956 (2013.01); B60W 30/143 (2013.01); B60W 30/18163 (2013.01); B60W 60/0027 (2020.02); G01C 21/3461 (2013.01); G06N 3/08 (2013.01); G08G 1/166 (2013.01); G08G 1/20 (2013.01); B60W 2420/403 (2013.01); B60W 2420/408 (2024.01); B60W 2520/00 (2013.01); B60W 2554/4041 (2020.02); B60W 2556/65 (2020.02); G08G 1/22 (2013.01)] | 21 Claims |
1. A real-time machine perception system for vehicle actuation of a first vehicle in a multi-vehicle environment, comprising:
an optical sensor for receiving a reflected portion of light from one or more objects within a field of view, and generating an optical signal based on the reflected portion of light;
an optical processing component coupled to the optical sensor, the optical processing component comprising:
an optical neural network (ONN) configured to:
detect the first vehicle is approaching a conflict zone;
in response to detecting the first vehicle is approaching a conflict zone, determine weights for a deep learning algorithm by solving a non-linear control problem involving multi-vehicle merging or multi-vehicle synchronization for the first vehicle moving into the conflict zone; and
apply the deep learning algorithm to the optical signal, the deep learning algorithm having the weights determined by solving the non-linear control problem to the optical signal; and
an electronic actuation signal generation component coupled to the ONN configured to generate an electronic actuation signal for controlling an actuation system of the first vehicle;
wherein when the reflected portion of light is at least partially reflected off an obstacle in the field of view, the electronic actuation signal corresponds to an actuation of the first vehicle for the purpose of at least one of: (i) adjusting a relative velocity between the first vehicle and the obstacle; or (ii) collision avoidance between the first vehicle and the obstacle.
|