The behavior of robots in unprecedented conditions such as those encountered in a granular environment or fluid is not adequately understood. Inspired by the energy-efficient locomotion and planning strategies observed in  biological organisms, we develop limbless and legged robots to discover the principles of motion under the dominating influence of contact, fluid, or granular material.


We rely on both physical and computational models to advance research in robotics. The computational analysis relies heavily on high performance computing. For instance, we have been studying the influence of collision on limbless locomotion by conducting a massive amount of simulations using high-performance computing—equivalent to hundreds of years on a personal computer. In parallel, we have been developing physical models for the purpose of verification and validation.

The picture illustrates the Rattlesnake (sidewinder) marks on the dune in Death Valley California. By developed a computational and physical models to mimic this locomotion, we have been trying to understand the influence of topology, locomotion, and granular structure.   
The picture illustrates a candidate robotic snake used to to characterize the interaction with soil and investigate the motion behavior as influenced by topology and composition.
Pattern formation in snake trajectory while interacting with a pegs array. The color legend illustrates the distribution of the trajectory. The picture illustrates the result of approximately 1000 simulations.