A highly efficient weak target line-spectrum detection algorithm

To detect underwater weak targets, we propose an efficient line spectrum detection algorithm based on adaptive line spectrum enhancement (ALE) and frequency-domain block processing. The proposed algorithm is termed as normalized frequency-domain block least mean square (NFBLMS) algorithm, which is not affected by the input signal power in the weight iteration process. Through theoretical analysis and numerical simulation, it is shown that compared with existing algorithms, NFBLMS algorithm can implement ALE in real time and obtain higher system gain, furthermore, NFBLMS algorithm is more robust to step size, thereby producing a tradeoff between the convergence speed and the steady-state error. Therefore, NFBLMS algorithm is more suitable for the engineering application.