A multi-objective trajectory planning method and system for unmanned ships based on inverse reinforcement learning

Abstract

The present invention provides an unmanned ship multi-target trajectory planning method and system based on inverse reinforcement learning. The method includes: using reinforcement learning to obtain the optimal strategy pool, inputting the information of the final target state, and obtaining the optimal strategy pool to reach the final target point. Optimal path, control the unmanned ship to follow the optimal path; when there is an obstacle ahead, use inverse reinforcement learning to obtain a path that can avoid obstacles based on multiple target points, and control the unmanned ship to reach a new target point in stages , to achieve emergency obstacle avoidance. The system includes an initialization module, a strategy estimation module, a strategy optimization module, and a multi-objective point module. The beneficial effect of the present invention is that not only the global path planning can be realized, but also the calculation time can be reduced by using the trained strategy pool and multiple target points in complex sea conditions, and the emergency dynamic obstacle avoidance can be realized.

Description

technical field [0001] The invention relates to the field of unmanned ship path planning, in particular to a method and system for unmanned ship multi-objective trajectory planning based on inverse reinforcement learning. Background technique [0002] Human beings have never stopped exploring the earth. With the rise of artificial intelligence, various unmanned devices have been put into use, such as unmanned vehicles and drones. The use of these unmanned devices facilitates human exploration of more unknown areas. The ocean occupies 70% of the earth's surface area. How to explore the ocean and develop marine resources has become the focus of attention of all countries. Under the general environment of artificial intelligence, the research of unmanned ship is put on the agenda. [0003] Compared with the use of unmanned vehicles on land, the complex marine environment has brought many new challenges to the research of unmanned ships, such as submarine vortexes and underwate...

AI Technical Summary

Problems solved by technology

[0003] Compared with the use of unmanned vehicles on land, the complex marine environment has brought many new challenges to the research of unmanned ships, such as submarine eddies and underwater organisms. The movement of the manned ship creates difficulties

The path planning of unmanned ships is the key technology for the safe driving of unmanned ships. In some complex marine environments, traditional path planning algorithms are difficult to deal with these problems.

[0004] The Chinese patent applications with patent numbers CN201810229544, CN201811612058, and CN201910494894 involve the trajectory planning of unmanned ships, but in general, there are the following problems: First, the existing technology needs to know the obstacle information on the map in advance , there is no way to avoid the sudden appearance of dynamic obstacles; second, the existing technology sets a single target point, and once it encounters a large target point such as a submarine vortex or undercurrent, it cannot solve the next step planning problem; third, the existing The technology is mainly aimed at global path planning, and it is easy to fall into a local optimum and cannot cope with emergencies

Images