Analysis and testing of PSO variants through application in EEG/ERP adaptive filtering approach
- First Online:
- Cite this article as:
- Ahirwal, M.K., Kumar, A. & Singh, G.K. Biomed. Eng. Lett. (2012) 2: 186. doi:10.1007/s13534-012-0071-x
- 259 Downloads
An improved method for adaptive noise canceller (ANC) is proposed for electroencephalography (EEG)/event related potential (ERP) filtering in case of EEG self interference. ANC is implemented through five versions of Particles Swarm Optimization (PSO) technique.
A comparative study of the performance of PSO and its different versions such as constant weighted inertia PSO (CWI PSO), linear decay inertia PSO (LDI-PSO), constriction factors inertia PSO (CFI-PSO), nonlinear inertia PSO (NLI-PSO), and dynamic inertia PSO (DI-PSO) has been done. Fidelity parameters like signal to noise ratio (SNR) in dB, correlation between resultant and template ERP, and mean square error (MSE) are observed with varying range of particles and inertia weights.
In this the results of two data sets, simulated ERP and real visual evoked potential (VEP) are compared. Fidelity parameters as well as quality (shape) of extracted ERP are determined with Kurtosis and skewness measures.
From the simulation results and comparative studies, it is found that that NLI and LDI version of PSO are most suitable for ANC for ERP filtering.