Estimation of pathological tremor from recorded signals based on adaptive sliding fast Fourier transform

Abstract

Pathological tremor is an approximately rhythmic movement and considerably affects patients’ daily living activities. Biomechanical loading and functional electrical stimulation are proposed as potential alternatives for canceling the pathological tremor. However, the performance of suppression methods is associated with the separation of tremor from the recorded signals. In this literature, an algorithm incorporating a fast Fourier transform augmented with a sliding convolution window, an interpolation procedure, and a damping module of the frequency is presented to isolate tremulous components from the measured signals and estimate the instantaneous tremor frequency. Meanwhile, a mechanism platform is designed to provide the simulation tremor signals with different degrees of voluntary movements. The performance of the proposed algorithm and existing procedures is compared with simulated signals and experimental signals collected from patients. The results demonstrate that the proposed solution could detect the unknown dominant frequency and distinguish the tremor components with higher accuracy. Therefore, this algorithm is useful for actively compensating tremor by functional electrical stimulation without affecting the voluntary movement.

Upper Limb Bionic Orthoses: General Overview and Forecasting Changes

“…For this reason, different types of filtration are used. The most common of them are: Hilbert transform [141], Fourier transform [142], and continuous wavelet transform [143,144].…”