Prosthetic robotic knee joint controller

Objectives:
- Capture the patient's intended knee motion

- Maintain less than 10% output error.

I led the development of a prosthetic robotic knee joint controller for patients with above-knee amputation, designed to accurately interpret user intent and translate it into precise joint motion. The project focused on capturing the patient’s intended knee angle from upper-thigh muscle activity and reflecting it in a robotic prosthetic knee with high fidelity. The system integrated electromyographic (EMG) sensors to acquire muscle activation signals and employed embedded control loops and advanced filtering algorithms to suppress noise and improve signal reliability.

The resulting controller achieved less than 7% error between user input and actuator output—exceeding the design requirement of 10%—while reducing overall system response time by at least 30%, demonstrating improved accuracy, responsiveness, and clinical viability.

Top collaborators: Joel Biju Thomas, Chenxin Zheng, Fangren Xu, Kexuan Yan