Adaptive Motion Planning And Learning Methods For The Study Of Autonomous Mobile Platform

The autonomous mobile platform is the ocean submarine, on land and air, autonomous mobile robot drones collectively. Now the country attaches great importance to marine development and utilization, and ocean military, energy development, resource exploration,etc. have been used in the different autonomous mobile platform. This paper is mainly focused on the unmanned underwater vehicle (unmanned underwater vehicle, the UUV) of adaptive motion planning problem, combined with the artificial potential field method and reinforcement learning method, carried out on UUV in environmental planning and study for an autonomous mobile platform motion adaptive planning and learning methods has important theoretical and practical significance..First of all, This paper summarizes the research situation of UUV motion planning in the environment and the common method, the artificial potential field method dynamic programming which has a minimal value and target accessibility issues proposed to give appropriate deflection angle improved method, through the simulation experiments verify the feasibility of the improved method..Secondly, for the static environment of UUV motion planning, the introduction of the reinforcement learning method based on static environment, a model was established. By using the proposed artificial potential field method and reinforcement learning method,combined with observations of the environment information division of the state space and action space, the design of reinforcement function. Through simulation experiments, it is proved that the motion planning of UUV in the static environment is adaptive, and the ability of self-learning.Thirdly, for the dynamic environment, the introduction of the "prediction" idea and design the obstacle avoidance layer and increase the safety campaign of UUV. In reinforcement learning method introduced eligibility traces, the eligibility trace), improved reinforcement learning method for reliability allocation problem. Through the simulation experiments, the study on the planning and learning methods of UUV in dynamic environment is achieved.Finally, through the UGV practicality and Simulation of UUV platform complete autonomous mobile platform motion adaptive test plan, design on enhanced learning obstacle avoidance test and four rooms room environment simulation test of boxes and other obstructions to verify the effectiveness and feasibility of the artificial potential field and reinforcement learning combined with the motion rules of the cost-effective method. The experimental results show that the developed planning system, autonomous mobile platform can be completed on the environment continue to strengthen learning and get better adaptive motion planning.