Multi-unmanned aerial vehicle motion planning method based on artificial potential field method and MADDPG

Abstract

The invention discloses a multi-unmanned aerial vehicle motion planning method based on an artificial potential field method and MADDPG. According to the method, high-quality experience of successfully planning a plurality of unmanned aerial vehicles to a target through the artificial potential field method is added on the basis of the original multi-unmanned aerial vehicle exploration environment experience, through the MADDPG algorithm training, samples are collected from exploration environment experience and high-quality experience at a certain probability, state information and environment information of each unmanned aerial vehicle serve as input of a neural network, the speeds of the multiple unmanned aerial vehicles serve as output, training of a motion planning strategy is completed, the multi-unmanned aerial vehicle autonomous obstacle avoidance in a complex environment is realized, and the target position is safely and quickly reached. According to the method, the Q values of the multiple unmanned aerial vehicles in different states and different actions are fully learned, the robustness of the strategy is improved, an excellent strategy with higher adaptability and higher flexibility is trained, and the method has a good application prospect in a scene of cooperative motion planning of the multiple unmanned aerial vehicles.

Description

technical field [0001] The invention belongs to the technical field of unmanned aerial vehicles, and in particular relates to a multi-unmanned aerial vehicle motion planning method. Background technique [0002] With the continuous development and improvement of science and technology, the multi-UAV motion planning technology, that is, the path planning technology that realizes the efficient obstacle avoidance of multiple UAVs, has been widely used in all aspects of human production and life. play an important role in revolutionizing industrial efficiency. The traditional multi-UAV motion planning method is dedicated to using some search algorithms and planning algorithms to calculate a better path when the environment is completely known and basically fixed, and the UAV positioning is accurate and meets the relevant motion trajectory. , such as A* algorithm, artificial potential field method (Artificial Potential Field, referred to as APF) and vector field histogram algori...

AI Technical Summary

Problems solved by technology

The public patent CN109059931A proposes a motion planning method based on multi-agent reinforcement learning. Based on the most basic reinforcement learning Q-Learning algorithm, the global agent and local agent with complementary functions are used to explore and utilize the flight environment, respectively. Corresponding to global path planning and local path planning, however, this method requires the action space to be discrete and the dimension cannot be too high, and the obtained strategy is discrete, which has poor adaptability in a dynamic and complex environment; while achieving efficient continuous decision-making, As an excellent multi-agent deep reinforcement learning algorithm, the MADDPG algorithm adopts the framework of centralized learning and decentralized action, without the need to establish realistic communication rules, and can solve the problem of non-stationarity of the environment well, providing an excellent solution for multi-agent reinforcement learning. basic framework

However, when this algorithm is applied to the motion planning of a large number of agents, there is a common problem that the dimension of the joint space is too large, and the training period is significantly longer, making it difficult to converge, and even the stability and dynamic adaptability of the training model are extremely poor. problems, it is difficult to apply to dynamic and complex environments, and has certain limitations

Images