Real-time prosthesis control using PID embedded control system

This project works to develop an intrinsic control system to autonomously operate a powered knee prosthetic leg. A custom powered knee was modified to include a 12-bit encoder, microcontroller, and motor driver. An embedded control system, employing a PID controller and created using MATLAB Simulink, controls the motor’s position. Five conditions for the PID constants were tested to establish the ideal gains for each angle transition. The angle transitions tested, which are analogous to points along a joint angle trajectory, were 10° and 40° flexion from neutral standing position and from 30° in extension to the standing position. The settling time, rise time, and overshoot were analyzed for each of the five conditions. For the 10° and 30° transitions, using lower P gains achieved values closest to the set target values, while slightly higher P gains accomplished this for the 40° transition. Higher P gains were required for transitions from lower to higher positions to counteract gravity. This approach could be used to drive the motor according to a predefined knee trajectory and establish specialized PID gains for each phase of the gait cycle to pair with real-time gait segmentation of IMU data for finite state control.