A study based on fourth-second order normalized cumulant in TDOA estimation

This work presents a problem of estimating the difference in arrival times of non-Gaussian signals added Gaussian noises at two separated sensors. The method is based on fourth-second order normalized cumulant and variable-step Least Mean Square algorithm. In this chapter, we investigate the normalized cumulants of the two different received signals by sensors to minimize the mean square error using variable-step Least Mean Square algorithm of time delay estimation. A new objective function is created in adaptive processing. When the adaptive filter converges, the weighted coefficient is calculated, while the value of time delay estimation is also obtained. Simulation results demonstrate that, under two different conditions of the noises, with the normalized cumulants, convergence curve performance via the proposed method is better than the method based on the fourth order cumulant.