MATLAB_Python_IMU_Demo

This work was performed at the University of Michigan Robotics Department with the Locomotor Control Systems Lab and the Neurobionics Lab.

Summary

This repository was developed to demonstrate a framework for combining MATLAB and Python for rapid control system prototyping. This particular example shows how to use MATLAB to perform sensor fusion with a low-cost IMU and to use the result to control the knee joint of the Open-Source Leg.

I presented this framework at the BioRob 2024 conference workshop entitled "Accelerating Wearable Robotics: Advancements in Mechatronic Prostheses and Exoskeletons and Rapid Control Prototyping" in Heidelberg, Germany. If possible after the conference, I will link the talk recording here.

Details

There are two main files:

1. IMU_Follower.m implements MATLAB's IMU Kalman filtering algorithm to calculate device orientation from raw accelerometer and gyroscope signals. We clamp the output between 0 and $\pi/2$ to prevent exceeding the hardware's range of motion. We convert this MATLAB function to a compiled library using MATLAB Coder (see IMU_Follower.prj).
2. BNO055_demo.py handles communication with the sensor and the Open Source Leg hardware. It leverages the OSL Software Library to control the knee joint of the prosthesis based on the calculated IMU angle.