Deep reinforcement learning reward function optimization method for unmanned ship path planning

Abstract

The invention provides a deep reinforcement learning reward function optimization method for an unmanned ship path planning. The method comprises the steps of S1, obtaining environmental information;s2, obtaining the distance between the unmanned ship and the obstacle and the distance between the unmanned ship and the target point; s3, giving a corresponding reward value according to the number of times that the ship arrives at the target point; s4, judging whether the ship is in a reward domain or not, and giving a corresponding reward according to a reward domain reward principle; s5, judging whether the unmanned ship collides with the obstacle or not, and giving a corresponding punishment value; s6, judging whether the ship is in the dangerous area or not, giving a corresponding punishment according to a dangerous area punishment principle, and otherwise, giving a reward according to a general situation reward principle. According to the method, the obtained rewards or punishmentsare increased or decreased by adding the reward domain near the target point of ship navigation, adding the danger domain near the obstacle and introducing the counting principle, the convergence speed of the deep reinforcement learning algorithm is increased, and the ship is instructed to avoid the obstacle more quickly to reach the target point.

Description

technical field [0001] The present invention relates to the technical field of path planning, in particular to a deep reinforcement learning reward function optimization method for unmanned ship path planning. Background technique [0002] At present, the economic ties of countries all over the world are getting closer and trade is more frequent. As an important means of transportation, ships play an important role. However, with the increasing density of ships and the increasingly complex navigation environment, maritime safety accidents occur frequently. The data in recent years shows that the main cause of shipwrecks is that the obstacles that the crew did not detect in time collided with the ship during the voyage. At the same time, in some cases, it is not suitable for a manned ship to go to the work place to perform tasks, and the ship needs to sail autonomously to cope with the complex and changeable harsh environment at sea, which requires the ship to have the funct...

AI Technical Summary

Problems solved by technology

[0008] In order to solve the problems of slow convergence speed and long training period of the traditional reward function, the present invention proposes a deep reinforcement learning reward function optimization method for unmanned ship path planning

Images