Flexible and Hybrid Action Selection Process for the Control of Highly Dynamic Multi-Robot Systems

This chapter presents a behavioral mechanism of control in order to break the complexity of multi-robot control systems. Specifically, this chapter proposes a Hierarchical Action Selection Process (HASP) which aims to coordinate a set of elementary controllers endowed in behavioral control architectures. This process allows at the scale of the robot to coordinate in a hierarchical and flexible way the activity of a set of elementary controllers, and at the scale of the group, the coordination of robots’ interactions for reaching global objective and desired mass effects. The performances of the HASP were improved via the addition of an appropriate mechanism of fusion of actions leading thus to a Hybrid-HASP. Two architectures of control using respectively the HASP and the Hybrid-HASP are proposed to achieve the cooperative box-pushing task. The validation of the proposed mechanisms of control was made through experimentations using minimalist mobile robots and more intensively according to statistical studies achieved on a large number of data gotten thanks to MiRoCo simulator.