This study aims to develop neural machine interfaces synthesizing high-density electromyography and deep learning to achieve accurate, reliable, and generalizable neural machine interfaces for wearable robots.
Neural interfaces, novel technologies to extract movement intention from muscle electrical signals (i.e., electromyography), have demonstrated great success in providing customized rehabilitation for people who have experienced a stroke and intuitive control of robotic prostheses for people with amputation. However, the existing neural interfaces are not stable/reliable for long-term use and require periodic time-consuming calibration, which significantly hinders the acceptance and application of neural interfaces in clinical populations. To address the challenge, this study will exploit artificial intelligence, specifically deep convolutional/recurrent neural networks (CNN/RNN), to capture reliable and generalizable features from high-density electromyography (HDEMG) signals recorded using advanced arrays of electrodes, to provide an easy-to-use, convenient, and generalizable neural machine interface.
The students will learn to record, process, and analyze electromyography using an advanced high-density EMG acquisition system (https://otbioelettronica.it/en/quattrocento/), implement and explore different deep neural networks for neural signal processing. The students will have the flexibility to work on subtasks that fall within the scope of study and align with their interests. It is desired that the students commit to work 10-20 hours/week for 15 weeks (January 8, 2024, to April 25, 2024).
Project Dates
Start Date: 1/8/2024 - End Date: 4/25/2024
Students Needed
Type of Project
Individual
Student Responsibilities
Here is a list of position responsibilities: 1. Attend project meetings 2. Conduct literature reviews 3. Implement control algorithms (in MATLAB or Python) for the robotic knee-ankle prosthesis 4. Conduct experiments to validate the performance of different control methods 5. Investigate the effects of different control parameters on joint kinematics and kinetics
Time Commitment
>10 hours/week hour(s)
Student Requirements
Preferred qualifications: 1. Basic knowledge of biological signals or machine learning 2. Programming skills in Matlab or Python 3. Experience in data collection and data analysis 4. Excellent written and verbal communication skills 5. Experience of working with Excel
Interested in Working With the Following Programs
For EXCEL URE Students Only
Additional Notes