PID Tuning for Vision Based Mobile Robots Formation Control

Abstract

Multiple autonomous robots are more and more used in fully automated warehouses where they can operate safely sharing the same space using a simple obstacle avoidance strategy and transport objects from one area to another. For transporting large or heavy objects, it is sometimes interesting to use several mobile robots cooperating physically and keeping the same geometric formation. In this case, the control of the robots’ formation is a crucial issue. In this context, our work proposes a strategy based on an ARTAG marker tracking for building and maintaining a Leader-Follower (LF) configuration using no other communication means than indirect communication via tracking. The tracking is using a PID (Proportional-Integral-Differential) controller for the following process and AR track Alvar vision system for the recognition and localization process. The proposed method only needs distance and angle errors information during the tracking. We defined the total error during the following phase as the performance measurement to be minimized. For PID tuning, the manual tuning method is used and compared with classical PID and conventional PID tuning methods including Ziegler-Nichols method, Tyreus-Luyben method and Cohen-Coon method. Experiments are validated and conducted on two real mobile robots, the master robot carrying the marker is manually controlled and the follower is keeping its position relative to the master using the AR tag tracking.