An Extended Kalman filtering of accelerometer and surface electromyography data for attenuation of pathological tremor

Previously a Kalman filter has been developed to estimate joint angle of the tremulous upper limb using data from accelerometer (ACC) and surface electromyography (sEMG). Results have shown that the fused information can be useful for actively compensating the tremor and helping the clinicians in tremor diagnostics. In this paper, an improvement for the current algorithm is proposed by implementing Extended Kalman Filter (EKF). There is electromechanical delay between the muscle activation (sensed by sEMG) and onset of motion (sensed by ACC). Thus some information from the sEMG will be extracted first then it will be fed to the EKF algorithm together with the measurement from ACC.Weighted-Frequency Linear Combiner (WFLC) is used to extract the frequency of the sEMG data. The EKF will then be able to estimate the amplitude and phase of the tremor, along with the accelerometer bias. The extracted parameters of the tremor will be useful for its attenuation. The recursive nature of WFLC and EKF algorithm enables a real time implementation.