Lobster-inspired hybrid modules for compliant robotic applications

The emerging field of soft robotics provides new insights into designing robots that can compliantly interact with humans, which is desirable for a variety of applications such as medical tools, rescue robots and exoskeletons. However, these flexible devices also present new challenges in sensing and controlling their continuous deformations.

Inspired by the hybrid structure of lobsters, a series of hybrid modules have been developed consisting of both soft and rigid components. The idea is that the soft chambers would be pneumatically actuated to generate safe and compliant actuation, and rigid exoskeletons can protect these chambers and provide convenience in both sensing and control. Also, different modules can be assembled to generate compilated motion profiles, and novel stiffness control strategies have also been developed. As a result, we can assemble a variety of compliant robots based on different requirements using these modules.

This project will further pursue developments in efficient deep learning for resource constrained devices.