Automatic function-isomorphic configuration recognition and control for UBot modular self-reconfigurable robot

摘要

Versatility of configurations is a significant feature for modular self-reconfigurable robots. For an arbitrarily assembled typical configuration, it is not easy to find the control method quickly in case of the symmetries of the configuration and module shape. To solve this problem, a database with lots of known configurations has been established. All the configurations in the database have their own control methods which have been studied before. Therefore, the way for matching the new configuration to one existing configuration in the database is quite important in locomotion and self-reconfiguration control of the self-reconfigurable robotic system. A combined configuration recognition method based on Matlab graphisomorphism function and improved three-dimensional linked list of modules algorithm was proposed. Firstly we use graphisomorphism function to find the configurations sharing the same underlying graph structure, then use improved three-dimensional linked list algorithm to acquire the ID and gait mappings. The method has been verified by simulation and experiment.