Variable step size p-norm-like constraint sparse underwater acoustic channels estimation method

Aiming at the high Bit Error Ratio (BER) of the class of norm constraint Normalized Least Mean Square (NLMS) algorithm in sparse channel estimation of Orthogonal Frequency Division Multiplexing (OFDM) underwater acoustic communications, a variable step size p-norm-like constraint channel estimation method is proposed. The method is based on the improved double Logistic function. Furthermore, it introduces the error signal autocorrelation function, thereby the step size and zero-attracting term are adjusted dynamically. Therefore, the convergence speed and estimation accuracy can be well compromised. Numerical simulation results show that the maximum performance improvement of the proposed channel estimation method is 72.3% faster convergence rate and 95.9% lower steady-state error in comparison to conventional ones. Compare to other channel estimation methods in the same category, the data processing results of the lake experiment demonstrate that the BER decreases by 2∼3 orders of magnitude under the shallow sea multipath sparse underwater acoustic channel, achieve zero-error underwater acoustic communication.