Automatic Tracking Method of Unmanned Vehicle Trajectory Based on Reinforcement Learning

Up to now, unmanned driving is still a challenging research field in academia at home and abroad. The vehicle trajectory tracking technology is a very critical and urgent link, because it provides important information for intelligent traffic monitoring. The reinforcement learning method is an important method for learning in an unknown environment. In the field of artificial intelligence machine learning, reinforcement learning research has made great progress in theory, algorithm and application, and has become a current hotspot in research. Unmanned vehicle trajectory tracking is one of the key technologies in the field of unmanned driving research. It uses built-in sensors to perceive the environment, uses trajectory planning algorithms to generate the required path in real time, and the decision system selects the best path. Finally, the built-in path tracking controller implement it. This article mainly adopts the experimental analysis method to discuss how to break through the problem of automatic trajectory tracking technology in the support of enhanced learning by unmanned vehicles, and compare and analyze the expected yaw rate and actual yaw rate and frequency of the target vehicle. According to the experimental research results, the expected yaw rate and actual yaw rate of the unmanned vehicle trajectory automatic tracking test are relatively close, and the test system in this test has a certain tracking effect.