Authors: Alberto Miranda, Samannoy Ghosh, Yong Lin Kong
Mentors: Yong Lin Kong
Insitution: University of Utah
Microscale robots can impart a broad range of functionalities in the biomedical domain that can be leveraged to address unmet clinical needs, including noninvasive surgery and targeted therapies. Conventional robot navigation methods typically involve specific gaits suited for certain environmental conditions. However, implementing the same conventional methods inside a human body is highly challenging. As the human body is a complex and dynamic environment, a microrobot must adapt to these complex and challenging environments to perform targeted studies. Previous research demonstrated an integration of an untethered, 3D-printed three-linked-sphere crawler with a model-free reinforcement algorithm. The work done with the theoretical Najafi–Golestanian three-linked-sphere mechanism was its first experimental integration with a reinforcement learning algorithm as a relatively simple and highly scalable self-learning robot that can navigate in unconfined and confined spaces. The progress presented in the current research is a direct continuation of the previous work on the 3-linked-sphere crawler. While the previous work focused on developing a proof of concept for adaptive gait learning for the crawler, the current work focuses more on the challenges of implementing the robot in a low Reynolds number fluid medium. Our current research hypothesizes that a self-learning autonomous system could demonstrate successful gait adaptation in a low Reynold’s flow environment. The design of our robot has been significantly improved to make it sustainable for extended use under viscous fluids. The research presented outlines the work that has been done to transition the robot from a crawler into a swimmer, the challenges that have been faced, and how they have been addressed. Successful implementation of this 3-sphere-swimmer will be a step forward in integrating machine learning tools into microswimmers for autonomous gait adaptation inside the human body.