LEARNING SYSTEM AND LEARNING METHOD FOR OPERATION INFERENCE LEARNING MODEL FOR CONTROLLING AUTOMATIC DRIVING ROBOT

摘要

Provided is a learning system 10 that is for an operation inference learning model 70 for controlling an automatic driving robot 4, carries out reinforcement learning of the operation inference learning model 70, and comprises the operation inference learning model 70, which is for using a travel state of a vehicle 2 including the vehicle speed to infer operation of the vehicle 2 that will result in the vehicle 2 traveling according to a specified command vehicle speed, and the automatic driving robot 4, which is mounted in the vehicle 2 and causes the vehicle 2 to travel on the basis of the operation, wherein advance reinforcement learning of the operation inference learning model 70 is carried out through the application, to the operation inference learning model 70, of a simulated travel state output by a vehicle learning model 60, and after the advance reinforcement learning has finished, additional reinforcement learning of the operation inference learning model 70 is carried out through the application, to the operation inference learning model 70, of a travel state acquired by causing the vehicle 2 to travel on the basis of the operation inferred by the operation inference learning model 70.