| This work focuses on the design of warehouse multi robot system.Several key technical problems are studied: location navigation,path planning and multi robot task scheduling.The content of this work mainly includes the following aspects:Firstly,to deal with the problem of warehouse robots indoor positioning,a location method based on data fusion of visual landmark and odometer is proposed.With the purpose of acquiring positioning information,camera model is applied to obtain rotation and translation between landmark and camera.With the analysis on the angle error of gyroscope and odometer,a method of angle fusion based on variance weight is completed.It can deal with the low update frequency and discontinuous information of beacon positioning method.Kalman filter is used to fuse the information of odometer and visual positioning to improve the ability of single sensor.Secondly,classical A* algorithm is realized for robot path planning.A improved A* algorithm is also provided to reduce the number of turning,through path corner limitation.One-way traffic is used to constraint the path to avoid collision in multi-mobile robot path planning.Finally,the warehouse multi-robot picking system is designed.Aiming at the congestion problem of multi-robot picking system,a multi-robot path planning method based on congestion control is proposed.This method improved efficiency by appending forecast congestion probability and congestion probability map to A*.A multi-robot task scheduling strategy based on greedy policy is designed and implemented through the evaluation function of task scheduling.Finally,A task simulation software based on state transfer is designed.Several simulation experiments are carried out to verify the effectiveness of the strategy in aspects of collision avoidance,path length and total run time. |
PDF Full Text Request