Formation control of multiple autonomous mobile robots with inaccurate sensor information

Abstract

This thesis develops a control approach to generic formation tasks of multiple mobile robots with inaccurate sensor information. Inaccurate sensor information means that all the robots have only local sensors that cannot accurately measure absolute distances and directions of objects. The main formation task in this paper is to generate formation in which the lead robot becomes the reference of formation, and the others determine their position in relation to the lead robot. The logic is developed considering generic situations and increasing the number of robots participating in formation. Petri nets are used for modeling and design of the control logic, which can visualize the control models, make it easy to check the states of each robot, and recover from errors. To evaluate the developed logic, physically homogenous mobile robots are designed and built. Each robot is equipped with eighteen infrared sensors and a UHF transceiver module. The developed control approach is demonstrated with experiments to be successful and efficient for the formation problems of autonomous mobile robots.