US 12,259,735 B2
Deep model reference adaptive controller
Girish Chowdhary, Champaign, IL (US); Girish Joshi, Redmond, WA (US); and Jasvir Virdi, Urbana, IL (US)
Assigned to THE BOARD OF TRUSTEES OF THE UNIVERSITY OF ILLINOIS, Urbana, IL (US)
Filed by THE BOARD OF TRUSTEES OF THE UNIVERSITY OF ILLINOIS, Urbana, IL (US)
Filed on Jun. 22, 2021, as Appl. No. 17/354,912.
Claims priority of provisional application 63/042,996, filed on Jun. 23, 2020.
Prior Publication US 2021/0405659 A1, Dec. 30, 2021
Int. Cl. G05D 1/10 (2006.01); G05B 13/02 (2006.01); G05D 1/00 (2006.01); G06N 3/04 (2023.01); G06N 3/08 (2023.01)
CPC G05D 1/101 (2013.01) [G05B 13/027 (2013.01); G05D 1/0088 (2013.01); G06N 3/04 (2013.01); G06N 3/08 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A device comprising:
a processing system including a processor; and
a memory that stores executable instructions that, when executed by the processing system, perform operations, the operations comprising:
receiving sensor data from one or more sensors of an autonomous vehicle;
determining based at least in part upon the sensor data, at a slower time-scale, inner layer weights of an inner layer of a deep neural network;
providing periodically to an outer layer of the deep neural network, directly from the inner layer, a feature vector based upon the inner layer weights;
determining, at a faster time-scale, outer layer weights of the outer layer, wherein the outer layer weights are determined in accordance with a Model Reference Adaptive Control (MRAC) update law that is based upon the feature vector from the inner layer, and wherein the outer layer weights are determined more frequently than the inner layer weights; and
using one or more estimates of uncertainty from the outer layer of the deep neural network to cause one or more actuators, one or more motors, or a combination thereof to control operation of the autonomous vehicle via real-time adaptive feedback.