A decentralized control framework for modular robots

Abstract

Distributed control paradigm offers robustness, scalability, and simplicity to the control and organization of module based systems. MSR (modular self-reconfigurable) robot is a class of robot that best demonstrate the effectiveness of distributed systems, as all modules in the robot are individuals that perform their own actuation and computation; the behavior of the complete robot is a collective behavior of all independent modules. In this paper, a general control framework, named general suppression framework, is proposed and a distributed control system based on the framework is presented. The control system is designed to control a set of MSR robots configured into a planar manipulator arm. All modules in the manipulator arm contain their own processing and actuation units, which allow them to evaluate and react to the environment independently. The modules can perform passive communication with their immediate neighbors and can exhibit aggressive or tolerant behavior based on the environment change to generate emergent group behaviors. A simulation program is developed to demonstrate the effectiveness of the distributed system in controlling the module based planar manipulator arm.