Distributed Kinematic Inversion Technique for Self-Reconfigurable Modular Robots

In recent years self-reconfigurable modular robots have witnessed an increasing interest by part of the robotic community. Despite recent robots exhibit advanced electromechanical designs, in the field of supporting control techniques, strong results have been registered only within locomotion, while the manipulation capabilities of existing systems are still quite limited. Aiming to provide a contribution along such a direction, the present paper proposes a computationally distributed control technique allowing to control the motion of the end-effector of any considered modular kinematic chain in its operational space. The proposed strategy is not based on explicit roles assignment: all the modules are regarded in the same manner and can therefore be on-line inter-exchanged on wish. Each module just knows its physical parameters and is totally unaware of the characteristics of the overall chain (including the number of its d.o.f.'s). All the information needed for coordinating purposes are on-line obtained through data exchanges with the other modules. In this way a global self-coordinating behaviour is autonomously established along the chain by solely exploiting the control and communication capabilities provided by on-board low-cost local processing units. No additional centralized external control hardware with global knowledge of robot geometry or kinematics is necessary. Just a simple MMI for sending high-level commands and monitoring their execution.