Research On Tracking Motion Control Method Of Multi-omnidirectional Mobile Robot

In recent years omnidirectional mobile robot has been playing an important role in various fields of society by virtue of its unique motion characteristics.With the continuous expansion of robot application field,it is difficult for a single robot to cope with the changeable working environment and complex tasks.In this case,multi-robot cooperation has gradually become one of the hot issues in scientific research.Through the combined application of multiple sensors,this paper studies the formation transformation of multi-omnidirectional mobile robots in the "leader-follower" mode and the tracking motion control of the robot formation.The main research content of this paper can be divided into the following parts:(1)Corresponding kinematic models were established for different motion modes of omnidirectional mobile robots,and the dual virtual wheel model was used to replace the original four-wheel steering model to simplify the calculation process.A motion model selection method was proposed and a model selection function was established to enable the robot to select different motion models at different speeds and turning radii,giving full play to the advantages of omnidirectional mobile robots.(2)The kinematics modeling of common line formation and triangular formation in the"leader-follower" mode is carried out,and the control structure of the leader robot and the follower robot is given respectively.Based on the virtual robot model,the formation control problem is transformed into the path tracking problem between the follower robot and the virtual robot.(3)Several key technologies for tracking motion control of multi-omnidirectional mobile robots are studied.In view of the problem that Hector SLAM algorithm cannot eliminate cumulative errors and has no map correction ability,this paper optimizes it to improve the accuracy of map construction;In view of the phenomenon that the vision sensor is easy to be blocked in practical applications,this paper proposes an extensible multi-robot information interaction positioning scheme,which enables the follower robot to complete its own positioning work by information interaction with other follower robots.In this paper,the path planning method of robot formation is improved.The formation changes are incorporated into the RRT algorithm to give the formation the ability to pass through narrow areas.At the same time,the cubic B-spline curve is fitted with the tortuous part of the planned path,making the final path smoother.(4)The path tracking control problem of robot formation is studied in this paper.The path tracking problem of the leader robot and the follower robot is transformed into the problem of eliminating the position and attitude errors of the respective target points.A formation path tracking controller is designed based on Backstepping control method and motion model selection method proposed in this paper.According to the control algorithm proposed in this paper,a layered control system based on hybrid control architecture was built,and a ROS wireless communication system based on Wi-Fi wireless communication network was designed,which enhanced the coordination of formation motion.Finally,this paper verifies the feasibility and practicability of tracking motion control of multiple omnidirectional mobile robots in the actual scenario by using the method of omnidirectional mobile robot prototype.