Enhancement of Signal to Noise Ratio using Bispectrum: ABR Signal Estimation Application

Bispectrm has been widely used to enhance the SNR. This is founded on the assumption that the HOS properties of the signal of nterest are different to the HOS properties of the measurement noise. In the present work, we consider the use of Bispecturm techniques when repeated measurement are made of a deterministic signal embedded in random noise where SNR is the range of OdB down to -17 dB. The performance of estimators and reconstruction algorithms are evaluated using simulated evoked potential data. We conclude theat the best performance is achieved if we reconstruct the phase by bispectrum and amplitude through a secound order method such as Spectral Subtraction.