An Intelligent Unmanned Control Method for Redunant Moving Agent

The paper is about redunant moving agent with intelligent unmanned control. The main task is to complete the construction of redundant fault-tolerant control system of deep neural network, including the infrastructure construction of unmanned agent simulation, the initialization of agent parameters, the construction of redundant controller, and the construction of reinforcement learning decision model. The main purpose is to generate simulated floating point data to train the model, including designing the expected rate and path, kinematics simulation, and training data generation. The kinematics simulation scene construction and decision-making model training use deep learning, whose effect of the system performance is significant.