Distributed Control For Multiple Robots Based On Feedback From Wireless Network

The distributed control of multiple robots is one of the hot researches in the word. The leader-follower construction of formation is used in this paper, and the virtual leader is put forward to realize the distributed structure. The strategy of rounding up the dynamic target with multiple robots can be departed in two: 1.virtual leader tracks dynamic target; 2. follower robots plans their path with the relative position to virtual leader and the planned position of virtual leader. Follower robots keep their cooperation by holding the relative position to virtual leader. All robots do the path planning and trajectory tracking themselves, and the rounding of the target is accomplished by all the robots. The research work of this paper is departed into four aspects:Firstly, to the control platform of multiple robots, this paper builds a network control system for multiple robots. The system frame and work process is introduced in detail. The system is formed of monitor layer, communication layer and object layer, and their hardware resources and functions are explained and analyzed.Secondly, the success of trajectory tracking for a single robot is the base of rounding the dynamic target multiple robots. The kinematics equation of the robot is analyzed, and the error model of trajectory is built and linearized. Then the linear model is used to designed LQR and PID controllers. The simulations on MATLAB of two methods both show the effectiveness, then the LQR controller is selected after their performance are compared.Thirdly, the prediction of target’s trajectory is a key step in the rounding of the target. This paper adopts the discrete Kalman filter to predict the trajectory of the target. The current position is used in the discrete Kalman filter to accomplish the state filtering. And then the state after filtering is used in position prediction. And then, the numerical computation is employed to explain that as the change of the target’s speed is bigger or the length of prediction is longer, the error of prediction is bigger.Finally, the distributed control is used to achieve the rounding of dynamic target by multiple robot. The leader-follower structure based on centralized control is improved, and the virtual leader is used to construct the distributed formation structure. The relative position of rounding robots to virtual is obtained from the expected formation. According to the predicted position of target, the rounding robot plans the position of virtual leader. And then the rounding robot uses the planning position of virtual leader and its relative position to virtual leader to plan its own position, and modified its planning position based on the restraint of path planning. In the end, the rounding method is proved to be feasible by the result of experiment.