A Method for Noise Reduction to Differential Chaos Shift Keying

This paper proposed a method for noise reduction to DCSK communications systems using weighted least square support vector machine (WLS-SVM). By exploiting the chaotic dynamics involved, the short-term robust estimation of chaotic signals can be obtained by first using a support vector machine with a least squares cost function and, then associating weighted values to the error variables based upon the resulting error variables from the first step. Computer simulation for the noise performance show that the noise reduction technique improves the overall noise performance in contrast to theoretical noise performance of DCSK communications systems