An application of the Kalman filter for EEG/ERP signal enhancement with the autoregressive realisation

Electroencephalogram/Event-related-potentials (EEG/ERP) signals have lower amplitude, are popularly known for their nonstationary nature and are easily exposed to other surrounding and biophysical artefacts. In the present work, the first attempt has been made to enhance the EEG/ERP signals using metaheuristically optimised Kalman filter parameters. The adaptive Kalman filter (AKF) parameters on proper tuning provide stable filtering operation. This parameter tuning is achieved metaheuristically by the recently proposed social mimic optimisation algorithm (SMOA). Further, it is tested in various noisy environments and compared with benchmark optimisation algorithms such as particle swarm optimisation algorithm (PSOA) and symbiotic organism search algorithms (SOSA). The proposed methodology manifests better results and surpasses other techniques in the literature. Hence, the proposed method can be used to enhance the EEG/ERP signal.