CAREER: Controlling Undulatory Soft Robotic Swimmers using Embodied Intelligence

This Faculty Early Career Development (CAREER) grant will fund research that enables soft-bodied unmanned underwater vehicles that use the soft body itself for control, thereby promoting the progress of science, advancing the national prosperity and welfare, and securing the national defense. Unmanned underwater vehicles could become vital devices for monitoring the constantly changing coastal environment, supporting critical infrastructure at the bottom of the ocean, and defending our nation. However, existing underwater vehicles face challenges when operating in coastal waves and currents and under crushing pressures at the ocean’s depths. In contrast, fish can freely swim from the coast to the abyss by leveraging their soft body for high maneuverability, quiet swimming, and efficient deep-water operation. Robotic vehicles with soft bodies aim to harness these biological benefits; however, controlling these soft vehicles is a significant challenge. This project will attempt to solve this challenge by enabling soft robotic fish that harness the swimming benefits of biology and require minimal computing resources. To accelerate the use of biologically-inspired underwater vehicles and inspire an educated workforce, this project will integrate outreach and education into the research. This project will conduct outreach to coastal scientists to identify use-cases and requirements for bio-inspired underwater vehicles, will develop soft robot learning kits to spark scientific curiosity in 4th grade students, and will provide undergraduate research opportunities. This project aims to establish a fundamental framework to control undulatory soft robotic swimmers using the embodied intelligence of the swimmer’s own physical dynamics. It will achieve this goal using physical reservoir computing, which will turn a swimmer’s nonlinear fluid-structural dynamics into a physical embodiment of a neural network and use this embodied intelligence to control the undulatory gait. This project will combine physical soft robots with computational models to identify the relationship between the embodied undulatory gait and the swimming performance as a function of the body dynamics, muscle actuation, sensory feedback, and active body stiffness. This research involves three key objectives: 1) identify the fundamental relationships governing control of the muscles and undulatory gait using embodied intelligence, 2) determine the relationship between active body stiffness and embodied control, and 3) harness the identified relationships to control the soft robot across swimming conditions. Fundamental relationships that could be uncovered in this project will generalize to advance the control of soft robots using physical reservoir computing and will extend beyond the boundaries of control, morphology, artificial muscle, and sensor type. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.