Abstract | In this paper, we compare the expressions for the steady-state mean-square error (MSE), the optimum step-size, and the corresponding minimum tracking MSE of the normalized sign regressor least mean square (NSRLMS), the normalized sign regressor least mean fourth (NSRLMF), and the normalized sign regressor least mean mixed-norm (NSRLMMN) algorithms for the case of real-valued data. The expressions for the steady-state MSE, the optimum step-size, and the corresponding minimum tracking MSE of the NSRLMF and NSRLMMN algorithms based on energy conservation relation approach are available in the literature for the case of real-valued data. Thus, in order to compare these three algorithms, we have derived the expressions for the steady-state MSE, the optimum stepsize, and the corresponding minimum tracking MSE of the NSRLMS algorithm based on energy conservation relation approach for the case of real-valued data. Finally, simulation results to substantiate the analytical results of the NSRLMS algorithm are also presented for the case of real-valued data. |
---|