Locomotion systems (Multimo-Bat) | Physical Intelligence - Max Planck Institute for Intelligent Systems

The overal design of Multimo-bat. (a) Overall scenario of integrated three motions: jumping, gliding, and perching. The robot jumps from the ground by compressed leg mechanism; the leg mechanism become a wing with a membrane in the air; The robot clings on the wall with additional passive foot mechanism; The robot jumps on the wall for second jump and gliding. (b) Robot prototype. A red arrow denotes moving direction of the robot. The leg-wing mechanism can be rotated to change modes between jumping and gliding motions.

Hyun Gyu Kim, Matthew Woodward

Mobility in unstructured environments is a significant challenge for robotic systems; however, there are systems capable of operation in these environments. For example, flying squirrels, frogs, and snakes are able to jump from a tree and glide to move to other trees. Wood peckers, are able to fly and perch on the bark. These integrated motions provide a maneuverability to avoid danger from predators or obtain food. Even though technology does not allow for replication, with the correct level of abstraction, these systems can inspire designs which can improve the performance of their robotic counterparts.

Inspired by Desmodus Rotundus (common vampire bat), who is able to integrate jumping, flying and perching motions by using only one structure, we aim to develop a robot which integrates jumping, gliding and perching motions with minimal necessary components. Particularly, we analyze not only this biological system which demonstrates the desired locomotion modes, but also significant levels of integration between the modes. The integration concepts are then abstracted from this organism and applied to the development of a robotic system. The robot has a leg for jumping by storing energy in leg muscles. The leg also functions as a wing with a membrane on its top. The addition of a foot allows for clinging to vertical surfaces. This motion integration creates the potential to improve the mobility of robots in unstructured environments by minimizing the necessary components. These minimal components can result in a smaller, lighter, and higher performance system than that of a system which combines these motions independently.

In summary, the benefit of our approach is that employing multiple locomotion strategies can significantly improve the mobility of systems operating in unstructured terrain. By utilizing an integrated approach for the addition of locomotion modes, the performance of individual modes can be preserved while reducing the additional structure and actuation required, therefore, improving overall system performance.