Multi-robot Formation And Cooperative Confrontation Research

In recent years,with the continuous development of automatic control technology,communication technology and robot technology,robots have become more and more widely used in human life.Under certain application backgrounds,a single robot has great defects in terms of perception range,computing power,and anti-jamming capability.But multiple robots can cooperate to obtain a larger perception range,higher computing power and better anti-interference ability.Therefore,multi-robot collaboration has become a new direction in the development of robotics.In the multi-robot collaborative cooperation problem,the multi-robot formation problem is the basic problem of studying multi-robot collaborative cooperation,and the multi-robot cooperative confrontation is an extension of the multi-robot formation problem.At present,the research on multi-robot collaboration mainly focuses on simulation,and lacks verification on physical robots.This article uses the NAO robots of Japan's Soft Bank as the carrier to study the multi-robot formation and cooperative confrontation.The main research content is multi-robot formation,NAO robot monocular visual positioning and multi-robot hunting.For the research of multi-robot formation,the multi-robot formation algorithm was designed,and the details of multi-robot formation algorithm were elaborated,including designing the formation model,multi-robot communication with each other,calculating the expected position of following the robot and adjusting the posture and orientation of the robot to the desired position to complete the formation.And through the MATLAB simulation and NAO robot formation experiment,the feasibility of the algorithm is verified.In the aspect of monocular visual positioning of NAO robots,the NAO robot target recognition and positioning technology is elaborated.The target pictures taken by the NAO robot are segmented,denoised,and feature extraction is performed to extract the required information in the target pictures.For the target object,a monocular vision target positioning model is designed,and the position information of the target object is obtained through calculation.A NAO robot positioning experiment was designed to verify the effectiveness of the positioning algorithm.In the research of multi-robot cooperative confrontation,a multi-robot cooperative confrontation algorithm is designed.First,the leader robot is set by the principle of distance priority.Second,the leader robot calculates the position of the enemy robot relative to the leader robot through the monocular visual positioning model,and converts the position to coordinates under the hunting coordinate system.Third,the leading robot predicts the position of the enemy robot based on the extended Kalman filter at the next moment,and calculates the corresponding hunting point of each hunting robot.Finally,the leading robot sends the desired hunting point to each hunting robot through robot communication,the hunting robot adjusts its posture and moves to the desired hunting point to complete the hunting.The feasibility of the algorithm is verified by MATLAB software simulation and NAO robot experiment.