EMG Amplitude Estimation with Adaptive Smoothing Window Length

Typical EMG amplitude estimators use a fixed window length for smoothing the amplitude estimate. When the EMG amplitude is dynamic, varying the smoothing length as a function of time can produce a higher quality amplitude estimate. This paper develops and investigates (in simulation and experimentally) a new technique for adaptive window length estimation. The simulations suggest that the "best" adaptive filter performed as well as the "best" fixed-length filter. Both filter types had to be tuned to the conditions of the simulation. Experimentally, it was found that multiple channel EMG amplitude estimators consistently performed better than single channel EMG amplitude estimators. Results with the adaptive processor were inconclusive. Perhaps due to task difficulty, no differences in adaptive vs. fixed-length processors were observed when subjects were asked to use real-time EMG amplitude estimates (presented on a video screen) to track a rapidly moving random target. When the target speed was slow, the experimental results were consistent with simulation predictions.