Research On Technology Of Omni-directional Mobile Robot With Multi-sensor Fusion Based On EtherCAT

The de velopm ent of eco nom y and soc iety and the impr ove ment of pr odu-ctivity promote the application of robots more and more widely.With the progress of sensor technology,the robot system has a more powerful detection and perception ability,which greatly promotes the development of technology of robot application.However,in the process of autonomous navigation,most mobile robots' data sources rely heavily on a single sensor,and there are problems such as inconsistent bus protocols,poor real-time performance,and complex applications.Therefore,based on ROS platform and EtherCAT communication technology,sensors such as odometer of wheel,laser radar and depth camera were integrated in this project to build a mobile robot that can fully realize autonomous navigation function.Firstly,selection and design of the omnidirectional mobile robot are carried out according to its functions and technical indicators.The design of hardware and software is carried out respectively.The hardware and software systems are combined to build a physical platform,which is used as the experimental object for subsequent experimental verification.Secondly,the optimization of system is carried out for omnidirectional mobile robot.The ROS system and EtherCAT master station are integrated and applied to improve the real-time performance of the whole control system and the synchronization performance between motors.The servo motor is optimized by writing driver program,and the synchronization is verified by designing experiments.Then,there is a problem of complex calculation for the traditional RANSAC algorithm in the removal of mismatches.Through rough removal and fine removal of two steps to limit the number of iterations to accelerate the removal speed and improve the matching accuracy.Considering the positioning error caused by the wheel odometer is not accurate,the visual odometer is used to correct the wheel odometer,and both of them are fused to improve the positioning accuracy.Next,the point cloud data generated by the depth camera are transformed into laser-like data and fused with the two-dimensional laser data to obtain more comprehensive and rich environmental information.Because of the shortcomings of standard A* algorithm,the aim of improvement and optimization is achieved by improving heuristic function and extracting critical waypoints.The simulation experiment proves that the improved path length is shorter,the efficiency of reaching the target point is higher,and the dynamic window method and the improved A* algorithm have better effect of dynamic obstacle avoidance.Finally,according to the above research contents,the odometer position fusion,fusion mapping,navigation path optimization and navigation precision testing experiments are carried out respectively.