Dynamic obstacle avoiding method based on combination of neural network and Q-learning algorithm

Abstract

The invention relates to a dynamic obstacle avoiding method based on combination of a neural network and a Q-learning algorithm, and belongs to the technical field of robots. The dynamic obstacle avoiding method comprises the steps: initialization setting is conducted on relevant parameters, wherein initialization on parameters of the neural network and relevant parameters of Q learning are included; iterative training is conducted according to environment obstacle data and initialized parameters; according to current environment obstacle information, a current state of a moving robot is calculated and judged, a Q value is calculated and updated, and meanwhile parameters of the neural network are fed back and updated; according to the state after the parameters are updated, whether or notmovement of the moving robot is safe or not is judged; whether or not an iteration number reaches up is judged, and whether or not training continues is determined; and whether or not a target point is reached is judged, if not, Q table construction is conducted by using the neural network, a new round of iterative training is conducted, and if the target point is reached, navigation is ended. According to the method, the defects that the calculation time is long, and the convergence speed is low are overcome, and the obstacle avoiding efficiency in a dynamic environment is effectively improved.

Description

technical field [0001] The invention belongs to the technical field of robots and relates to a dynamic obstacle avoidance method based on the combination of a neural network and a Q learning algorithm. Background technique [0002] Path planning is one of the key elements of an autonomous mobile robot. It is hoped that the mobile robot can reach the destination as quickly and accurately as possible, and it is also required that the robot can safely and effectively avoid obstacles in the environment. At present, there are many better solutions to safely and effectively avoid obstacles and accurately reach the destination in a static environment. However, when there are moving obstacles in the environment, and the speed and position of the obstacles are changing all the time, the real-time and accuracy of the obstacle avoidance algorithm in the navigation process of the mobile robot are higher than those in the static environment. Higher, if you continue to use the algorithm ...

AI Technical Summary

Problems solved by technology

Each has its own advantages and disadvantages. For example, the artificial potential field method has a small amount of calculation and good real-time performance, but it is prone to local minimum points

In recent years, the Q-learning algorithm has been fa

Images