This paper presents computational and experimental evidence that it is possible to plan and execute dynamic motions that involve chain reconfiguration for modular reconfigurable robots in the presence of obstacles. At the heart of the approach is the use of a sampling-based motion planner that is tightly integrated with a physics-based dynamic simulator. To evaluate the method, the planner is used to compute motions for a chain robot constructed from CKbot modules to perform a reconfiguration, attaching more modules and continuing a dynamic motion while avoiding obstacles. These motions are then executed on hardware and compared with the ones predicted by the planner.
Scalable shape sculpting via hole motion: motion planning in lattice-constrained modular robots
