Coordinated Multi-robot Formation Control

A multi-robot system is formed when a group of robots interact with the environment as a single system. This system can also be in formation in order to accomplish tasks rather difficult or impossible to achieve with a single robot. This work presents results for a doctoral thesis research aiming the formation control of a mobile robot group in the active target tracking problem. All robots exchange information about their position (localization) in the world frame, and the target position and velocity forming a distributed system. Therefore, the main objective of this work is to search and track a target minimizing the total amount of uncertainty in the target’s observation, avoiding mates and obstacles, and obeying the characteristics imposed to the formation. A nonlinear model predictive controller was proposed in order to deal with the active target tracking problem through the formation of a multi-robot system. The formation controller created as the major contribution of this thesis is a nonlinear model predictive formation controller (NMPFC). This work also presents some contributions in the formation control concerning the obstacle avoidance, the temporary absence of the target and the formation control of a heterogeneous group (holonomic and nonholonomic robots). These contributions improved the formation making it versatile with dynamic uncertainties as well as the search and track of an initially absent target. Simulations and experiments with real robots were made, presented and discussed. Finally, this work is inserted in the FCT project PTDC/EEA-CRO/100692/2008 Perception-Driven Coordinated Multi-Robot Motion Control.