Research topics

A. Shape shifting (Self-reconfiguration)

1. Self-assembly		Reconfiguration from an arbitrary structure to a target structure (a big challenge)
2. Cluster flow		Parallel local reconfiguration forming global flow of module culsters
				movies
3. Sub-structure formation		Generation of small walkers
click figure to see movie ->
4. Metamorphosis		Small scale reconfiguration. e.g. a legged walker to a snake
				movies
5. Self-repair		Removal of a damaged module and reconfiguration (future work)
6. Re-connection		Re-connection of seperate robots (future work)
		movie

B. Locomotion

1. Locomotion pattern design and Experiment		Quasi-optimal pattern of multi-DOF motions were designed for various configurations of M-TRANs by offline calcuration by using locomotion controller based on CPG (Central Pattern Generator) network and optimizer based on GA (Genetic Algorithm). Obtained locomotion gait patterns were evaluated by experiments, where each gait pattern was played back.
many movies->
2. Locomotion control		Real time CPG control. Obtained quasi-optimal locomotion controllers, i.e., CPG network controller, were installed on modules' hardware. Distributed locomotion control was realized. Real time CPG control showed better performance than the play-back control above.
click figure to see movie ->	real time

C. Adaptation

Some studies has been made to realize the M-TRAN adaptive to the environmental changes. (Kamimura IROS2004)

1. Adaptive locomotion by CPG control	By the real time CPG control, locomotion gait is generated adaptively on the surface of variable friction.
click figure to see movie ->
2. Metamorphosis for adaptation to environment	When any module detects too large error from regulated motion, it initiates metamorphosis of the total body to adapt the environmental change and the robot start locomotion in another structure.
click figure to see movie ->

Top page