Real-Time Control of a Powered Prosthetic Leg Using Intramuscular EMG Signals
Abstract
Rationale: After a leg amputation, walking with an artificial leg (a prosthesis) is hard. It is especially difficult for people who lose their leg above the knee (so their ankle and knee are missing) because most prostheses are passive. This means that, unlike an intact leg, they do not provide power to help the person walk. Robotic prostheses that can provide power are being developed and may help people walk more normally with less effort, but controlling these devices is not easy. Onboard mechanical sensors provide information to tell the prosthesis what to do at each stage of walking, but the user cannot easily tell the prosthesis what they want it to do. Information from electrical signals (called EMG signals) that are produced when a muscle contracts, can be used to predict what the user wants to do. Combining leg muscle EMG signals with mechanical sensor information makes the control system work better and feel more natural. Unfortunately, recording EMG signals from the surface of the skin is difficult, as any change in the skin (e.g., sweating) or location of the electrode (which may vary from day to day when the prosthesis is taken off and put on again) affects the reliability of the signals. A control system that relies on varying signals may not work well and may cause falls, so users often have to recalibrate their control system with new EMG signals, which is inconvenient. One way to get more stable signals is to record EMG from inside the muscle. This can be done by implanting special sensors (MyoNodes) into leg muscles. MyoNodes can transmit EMG information wirelessly to a base station on the prosthesis. Objective: The overall goal of this project is to compare the control information provided by surface EMG to that obtained from inside the muscle (intramuscular EMG). The rationale is that EMG from inside the muscle will be more consistent and reliable than EMG recorded at the skin surface. This study will help us find out whether intramuscular EMG is better for controlling a powered prosthesis. Because the muscles that used to control the ankle are missing after a leg amputation, we will also surgically transfer nerves that used to control those muscles to new muscles in the leg. This is done in a technique called targeted muscle reinnervation (TMR). Once the nerves have grown into the new muscles, these muscles generate EMG when the person tries to move their ankle. We will use EMG signals from reinnervated muscles to control the prosthetic ankle and EMG from hip flexors (that lift the leg) and see whether this improves the person’s walking pattern and energy requirements when using a powered prosthesis. Specific Aims: Aim 1. Compare surface and intramuscular EMG signals in five individuals with unilateral transfemoral amputations before and after targeted muscle reinnervation surgeries. Aim 2. Compare real-time control of a powered prosthetic leg using EMG from MyoNodes and surface EMG. Aim 3. Incorporate intramuscular EMG signals into the mid-level controller of a powered leg prosthesis. Focus Area: We expect that using intramuscular EMG will make control systems for powered prostheses more reliable and consistent over time, which means that users will not have to recalibrate the system on a regular basis, and they will not need to have electrodes embedded in their liner or socket that may irritate their skin. We expect that more stable EMG signals, together with improved control of ankle function, will allow users to walk more efficiently with more normal walking patterns. Thus, this work addresses the Focus Area of “optimization of Warfighter performance following limb trauma or loss…intuitive control and sensation of prosthetic and orthotic solutions.” Who Will It Help: This research will benefit persons who have lost a leg above the knee and who wish to use a powered prosthesis and resume an active lifestyle. This may enable injured Warfighters to return to
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 10, 2021
- Source ID
- W81XWH2010892
Entities
People
- Levi J. Hargrove
Organizations
- Shirley Ryan AbilityLab
- United States Army