Multicollinearity problem of CPM communication signals and its suppression method with PLS algorithm

Abstract

The chaos phase modulation technique is one of the most promising solutions for underwater acoustic communications. Because the signal of this type modulates the phase of the carrier wave with pseudo-noise sequences, its demodulators are subjected to the distortions caused by the multicollinearity among the symbol vector elements. Because of this issue, this paper innovatively adopts partial least square regression into CPM communications to overcome the multicollinearity problem by modeling the relationships between the given predictor and response variables. The simulation experiments demonstrate that our method can provide a lower bit-error ratio compared to the reference algorithms.