Spring 2023 GRASP SFI: Wenzhong Yan, UCLA, “Mechanical Intelligence for Compliant Robots”

This is a hybrid event with in-person attendance in Levine 307 and virtual attendance via Zoom. This week’s presenter will be in-person as well.

ABSTRACT

Biological systems with deformable bodies exhibit remarkable abilities. However, compliant/soft robots still can not match the capabilities of their biological analogs in terms of adaptability, physical robustness, and autonomy partially due to the reliance on rigid, bulky silicon-based electronics. My research aims to address this challenge by creating self-sustained, autonomous soft machines through Mechanical Intelligence (MI) — embedding necessary functionalities into compliant materials and structures — with minimal or even no silicon-based electronics. The resulting material-like soft robots can behave like biological organisms, obtaining energy from the ambiance, moving on various terrains, and surviving extreme conditions, which may find applications in environment monitoring, rescue, exploration, human-machine interaction, education, etc.

In this talk, I will demonstrate the power of MI by achieving locomotion and autonomous interaction with the environment of soft robots in compliant origami materials and structures with high robustness to adversarial events, e.g., radiation and physical deformation. Firstly, I will show how to realize self-sustained oscillation by incorporating bistable mechanisms and conductive artificial muscles without discrete components or electronic control hardware. This oscillation can be used to generate locomotion for robots with only constant electrical power. Then, I will present an efficient method to rapidly design such oscillators from desired behavioral specifications, i.e. frequency. This method provides a powerful tool for facilitating the modeling, designing, and prototyping of such complicated dynamic compound systems. I will also introduce a method to achieve complete sense-decide-act loops in compliant materials for autonomous interaction with environments, demonstrated with several robots. I will close my talk with a preview of how to use MI to improve adaptability and convert/harvest environmental energy to power robots, toward achieving my ultimate goal of creating self-sustained soft robots to allow their widespread deployment into complex, extreme environments to perform challenging tasks.