Multi-Muscle FES Force Control of the Human Arm for Arbitrary Goals

IEEE Trans Neural Syst Rehabil Eng, 2014 · DOI: 10.1109/TNSRE.2013.2282903 · Published: May 1, 2014

Simple Explanation

This study presents a method for controlling a paralyzed human arm using functional electrical stimulation (FES). The goal is to achieve flexible motor outputs by controlling multiple muscles. The method involves using surgically implanted electrodes to stimulate muscles in the shoulder and arm. A model is created to map muscle stimulations to endpoint forces measured at the hand. The errors of the controller were characterized, and it was found that the total RMS error was 11% of the total range of achievable forces. The major error sources were random variability and model bias.

Study Duration

Not specified

Participants

One 54-year-old female with spinal cord injury

Evidence Level

Not specified

Key Findings

1
The magnitude of the total RMS error over a grid in the volume of achievable isometric endpoint force targets was 11% of the total range of achievable forces.
2
Major sources of error were random error due to trial-to-trial variability and model bias due to nonstationary system properties.
3
Nonlinear interactions between muscles made only modest contributions to the total error of the controller.

Research Summary

This study designs and evaluates a feedforward FES controller for flexible motor outputs, addressing challenges like decoupled control, redundancy, and nonlinear interactions in multi-muscle systems. The study quantifies total error, random error, model bias, and sources of model bias in multi-muscle force control, providing accuracy bounds for the skeletal system's muscle actuators. Nonstationary system properties were a large source of error, as re-estimating the model reduced the magnitude of the total RMS error by 2.5N.

Practical Implications

Controller Design

The findings suggest that FES controllers should account for non-stationary system properties and trial-to-trial variability, potentially through adaptive or robust feedback control strategies.

Simplification of Models

The minimal effect of nonlinear interactions between muscles simplifies FES controller design, as complex models of force production are not required.

Motion Control Strategies

Quantifying errors in isometric force control provides insights into the steady-state accuracy of torque actuators, which is useful in designing robot arm control strategies.

Study Limitations

1
The study only considered isometric force generation, which does not account for the nonlinear dynamics of the skeletal system.
2
The steady-state analysis did not consider the dynamics of the muscles themselves.
3
The study was limited by the amount of data available for these experiments.

Your Feedback

Was this summary helpful?

Back to Assistive Technology