IEEE Transactions on Instrumentation and Measurement · 2023 · 11 citations · 38 references
EngineeringBioengineeringBiomedical EngineeringBiomedical Signal ProcessingComputational AcousticsOptimizationBiomedical Signal FilteringDigital Signal ProcessingFrequency-domain AnalysisSignal ProcessingSystem IdentificationAdaptive FilteringAdaptive FiltersComputational EfficiencyEcho CancellationFrequency Domain ProcessingNonlinear System IdentificationRobustnessSignal Denoising
Frequency domain spline adaptive filters (FDSAF) have attracted a great deal of attention because of their enhanced computational efficiency. However, their performance will be extremely degraded in practical environments where non-Gaussian phenomena exist. In order to attain good robustness to non-Gaussian noises while ensuring a low computational burden, this paper proposes a frequency domain spline adaptive filtering algorithm based on the half-quadratic criterion (HQC), named FDSAF-HQC. Further, to offer a theoretical foundation for the proposed FDSAF-HQC, a detailed performance analysis based on the principle of energy conservation is conducted and then validated by simulations. Moreover, the computational complexity of FDSAF-HQC is evaluated and compared with other similar algorithms. Simulation results of nonlinear system identification corroborate the outstanding performance of the proposed FDSAF-HQC, i.e., better robustness and computational efficiency. The superiority of the proposed FDSAF-HQC is also supported by practical applications of echo cancellation and biological signal denoising.
38
Nonlinear spline adaptive filtering
Michele Scarpiniti, Danilo Comminiello, Raffaele Parisi et al. · Signal Processing · 2012 · 188 citations