A noise reduction method for discrete chaotic signals and its application in communication

This paper addresses the issue for filtering noisy contaminated discrete chaotic signals. A modified cubature Kalman filter combined with QR decomposition is proposed for estimating a nonlinear discrete system. Simulation results show the proposed method can effectively reduce noise on discrete chaotic signals. In comparison with other Kalman filters, the proposed method has a better filtering performance in the case of low signal to noise ratios (SNR), and has the similar performance in the case of high SNR. In addition, it is indicated by simulations that the proposed filtering method can also effectively implement chaos-based communication when message signals are modulated by additive chaos masking.