Monitoring of human neuromusculoskeletal system performance through model-based fusion of electromyogram signals and kinematic/dynamic variables

Abstract

This article demonstrates feasibility of utilizing model-based approach for monitoring the performance of human neuromusculoskeletal systems. The performance monitoring method utilizes an autoregressive moving-average model with exogenous inputs to link signatures extracted from electromyogram signals, to the forces produced by the limbs and limb velocities. This model is then utilized to quantify and track changes in the neuromusculoskeletal system dynamics over time. The changes were modeled using a measure of overlap between the distribution of 1-step-ahead model prediction errors observed at the beginning of the exercise, when the subject was rested, and 1-step-ahead prediction errors observed at any other time during exercise. As the subjects proceeded with their exercises and got increasingly tired, the distribution of modeling errors changed and diverged from the initially observed one, causing the overlap between the original and the most recent distributions of modeling errors, referred to as the Freshness Similarity Index, to decrease. Thus, Freshness Similarity Index could be used as a quantitative measure of the level of degradation the neuromusculoskeletal system performance through the exercise. The methodology was evaluated in two experiments, one related to an activity involving lower limb contractions, and the other involving jaw joint movements. In both cases, a decreasing trend in the Freshness Similarity Index clearly illustrated changes in neuromusculoskeletal system performance as exercise progressed. Furthermore, after rest, Freshness Similarity Index observed in both exercises recovered to their original levels, quantitatively and meaningfully showing that the corresponding neuromusculoskeletal systems of the two subjects indeed rested.