Coherent Accumulation Algorithm Based Multilevel Switching Adaptive Line Enhancer

For effectively extracting line spectrum signals from non-Gaussian noise, a coherent accumulation algorithm based two-path three-stage switching adaptive line enhancer (CATTSALE) was established. In the CATTSALE, one path is composed of a Fourth order cumulant variable step size sign coherent accumulation algorithm based adaptive line enhancer (FVSCAALE), a least mean square coherent accumulation based adaptive line enhancer (LCAALE), and a short-term correlation coherent accumulation based adaptive line enhancer (SCCAALE), whereas the other one is made up of the FVSCAALE and two SCCAALEs. Two-path may be switched by the on-off in the second level of the CATTSALE to obtain the much better effect of enhancing line spectra. Simulation tests with real data of the underwater moving target-radiated noise were carried out. The results have shown that the CATTSALE has immunity to Gaussian noise and that better ability to suppress non-Gaussian noise and to trace time-varying signals and outperforms the single LCAALE, the FVSCAALE, and the SCCAALE.