| In recent years,with the rapid development of robotics,manual operations have been gradually replaced by robotic operations in some traditional industries,and the application of multi-mobile robot systems will become the development trend of various industries in the future.Multi-robot systems often include problems,such as multi-mobile robot control system,task assignment,communication,path planning,etc.Among them,the research of path planning algorithm is a key part of the robot system.The goal of single robot path optimization is to plan a reasonable obstacle avoidance path from the starting point to the target point,while in the path planning of multi-mobile robots,avoiding collision and conflict between robots is a research point worthy of attention.This paper takes robot path planning as the main research object,designs optimization models and proposes solutions for different control systems,and realizes conflict-free driving between multimobile robots.The main research contents and conclusions include:(1)For centralized path planning,based on the improved A* algorithm,a coordination conflict strategy is designed to complete the path planning of multi-mobile robots.Considering the energy consumption factor model,for the problem that the turning angle is too large,a smoothing function and a guiding function are introduced to establish an improved A* algorithm model.At the same time,for the path conflict between multiple mobile robots,a coordination conflict strategy is designed,so that the multiple mobile robots can resolve the conflict and balance the energy consumption based on the initial path.Finally,the robot driving environment model is established by the grid method,and the Matlab is used to compare other algorithms for experimental simulation.The experimental results show that the energy consumption of the improved algorithm is better in the single robot environment;in the multi-mobile robot environment,the energy consumption of the improved algorithm is better.more balanced.(2)For distributed path planning,a mobile robot path planning based on reinforcement learning is proposed.Firstly,the characteristics of mobile robot path planning under reinforcement learning are analyzed,and the state space and reward function are designed to guide the robot to drive without conflict.The dynamic ε-greedy action selection strategy and the learning strategy of the Wo LF-PHC algorithm are introduced to make the robot quickly adapts to other robots while the number of its own pathfinding steps converges,and the frozen Q-value function strategy is introduced to solve the problem of different number of steps.Finally,simulation experiments are carried out on the Pycharm.From the experimental results,the proposed scheme can be conflict-free path finding for mobile robot,and balances the number of convergence steps and convergence speed,improves the path-finding efficiency of mobile robot under distributed reinforcement learning.(3)The distributed real-time collision avoidance algorithm is designed for the path conflict of multiple mobile robots in the wireless communication environment,and the feasibility of the distributed real-time collision avoidance algorithm design is verified.The proposed algorithm is a distributed multi-mobile robot collision avoidance algorithm based on the change of wireless signal strength.Different waiting times are designed for different robot numbers,and the backoff strategy is implemented.In order to further reduce the waiting time,a binary sequence priority collision avoidance algorithm based on RSSI is proposed.The transmission cycle is reassigned by converting the number into a binary sequence,and the next execution action is determined by comparing the transmitted signal with the received signal.Finally,the feasibility of the algorithm is verified based on the Webots.The experiments show that the proposed binary sequence priority method has shorter waiting time and higher efficiency than the fixed priority method. |
PDF Full Text Request