A modified time delay estimation algorithm based on higher order statistics for signal detection problems

Under the conditions of short data samples and low signal to noise ratio (SNR), practical applications for higher order statistics (HOS) are restrained by its high estimation variance and bad estimation accuracy. In this paper, a modified time delay estimation (TDE) algorithm based on HOS is presented, where normalization of input data based on second order statistics is adopted to take the place of calculation of self fourth order statistics of input signals in TDE. Therefore the resulting algorithm greatly reduces the estimation variance and computation time. Simulation results show that the proposed one has advantages over the traditional ones on both detection performance and computation efficiency.