Mobile robot fusion path planning method for urban environment

Abstract

The invention discloses a mobile robot fusion path planning method for the urban environment, and belongs to the technical field of mobile robot motion control. The method comprises the following steps that 1, a mobile robot judges whether global planning or local path planning is adopted according to obstacle information detected by a laser radar; 2, when the distance between the robot and the obstacle is greater than a set threshold value, an artificial potential field method is adopted to conduct a global path planning mode; and when the distance between the robot and the obstacle is smaller than a set threshold value, local obstacle avoidance is carried out by adopting a DDPG algorithm; and 3, the distance between the mobile robot and the target point is judged, and if the distance is smaller than a set threshold value, the task is completed; and if yes, repeating the first step and the second step until the target point is reached. According to the invention, the mobile robot can find a collision-free optimal path from the starting point to the terminal point.

Description

technical field [0001] The invention relates to an urban environment-oriented mobile robot fusion path planning method, belonging to the technical field of mobile robot motion control. Background technique [0002] The first prerequisite for a mobile robot to complete other tasks is to move to a designated target location, so path planning technology is one of the key technologies in mobile robot technology. In order to complete path planning for mobile robots in urban environments, not only static obstacles but also dynamic obstacles must be considered; due to the complexity and changeability of urban environments, it is very difficult to establish complete map information, so the unknown environment The path planning of the mobile robot under the environment is very important, but it also adds a lot of difficulty to the path planning of the mobile robot. [0003] At present, path planning algorithms can be divided into two categories: global path planning and local path p...

AI Technical Summary

Problems solved by technology

Rapidly-exploring Random Trees (RRT) is one of the classic algorithms based on random sampling. This algorithm can quickly realize path planning and is widely used, but it cannot guarantee the optimal path.

Although the artificial potential field method can be applied to unknown environments, it is easy to fall into local optimum

Representative algorithms based on swarm intelligence algorithms include genetic algorithms, particle algorithms, ant colony algorithms, and some fusion algorithms. Swarm intelligence algorithms are not suitable for robot path planning in unknown dynamic environments due to their high computational complexity.

[0004] In recent years, with the development of reinforcement learning and neural networks, deep reinforcement learning algorithms are considered to be the most potential method to solve unknown dynamic environments, including DQN (Deep Q-learning Network) algorithm, DDPG (Deep Deterministic Policy Gradient deep deterministic policy gradient) algorithm, A3C (Asynchronous Advantage Actor-Critic asynchronous advantage action evaluation algorithm), etc., deep reinforcement learning algorithm can realize end-to-end mobile robot path planning, without modeling the environment; but deep reinforcement learning algorithm There are problems such as slow convergence and instability that need to be solved urgently

Images