Unmanned aerial vehicle path planning method based on potential function reward DQN under environmental information unknown continuous state

Abstract

The invention discloses an unmanned aerial vehicle path planning method based on potential function reward DQN under environmental information unknown continuous state. The method comprises the following steps: firstly, establishing state space of an unmanned aerial vehicle in an environment, wherein the state space is a continuous state space and contains infinitely many states of the unmanned aerial vehicle; secondly, equally dividing 360 degrees into n angles and taking the angles as heading angles of the unmanned aerial vehicle, and establishing an action space of the unmanned aerial vehicle; then, calculating potential function reward of a target to the unmanned aerial vehicle and the potential function reward of an obstacle to the unmanned aerial vehicle, and superposing the rewardsand taking the superposed rewards as total potential function reward of the unmanned aerial vehicle; then, performing path planning training for a Q estimation network through the total potential function reward of the unmanned aerial vehicle; and finally, performing path planning under environment information unknown continuous state for the unmanned aerial vehicle through the trained Q estimation network. The method mainly solves a problem of path planning of the unmanned aerial vehicle without an environment model, satisfies requirements on state continuity of the environment, where the unmanned aerial vehicle is, when the unmanned aerial vehicle executes tasks; and the potential function reward accelerates path planning of the unmanned aerial vehicle, thus, the method has better applicability.

Description

technical field [0001] The invention belongs to the technical field of unmanned aerial vehicle path planning, in particular to an unmanned aerial vehicle path planning method based on a potential function reward DQN in an unknown continuous state of environmental information. [0002] technical background [0003] UAV path planning is a long-standing hot issue in the UAV field. It refers to the UAV planning an optimal or suboptimal collision-free path from the starting point to the target point under certain constraints. As the actual environment faced by drones becomes increasingly complex, planning a practical and effective flight path is a prerequisite for drones to successfully complete various tasks. The so-called path planning with unknown environmental information means that the UAV cannot predict the environmental information before planning the path, and the source of environmental information needs to be obtained by the airborne perception system. Since the UAV can...

AI Technical Summary

Problems solved by technology

However, these methods currently assume that the environment in which the UAV is located is a discrete grid map. Due to the limited state that this grid map can accommodate, the environment is required to be known, and the network convergence speed slows down, which is not very good. It satisfies the situation that the environment in which the UAVs are in the process of performing missions is unknown and the state is continuous

Images