A near-optimal noncoherent chaos-based communication scheme

In this paper we introduce a method for improving the performance of noncoherent chaos-based communication systems. This method involves transforming chaotic signals into predefined patterns according to the relative magnitudes of the signal samples. First, for large spreading factors, the transformation leads to a large correlation between two different chaotic signals having the same distribution. This property can be used to recover the information in a noncoherent system. If the chaotic map and the transformation function are known, we can generate another chaotic signal to construct the reference signal for use in the correlator. In this scheme, both the synchronization circuit and the reference-transmission are not required. This scheme is referred to as the near-optimal noncoherent (NON) scheme. Second, for small spreading factors where high autocorrelation of the transmitted signals cannot be achieved, we propose a near-optimal noncoherent scheme with transmitted reference (NON-TR). For comparison, the corresponding performance bound using coherent detection is also given. This reference scheme is called the near-optimal coherent (NOC) scheme. It can be shown that the NON scheme can achieve a similar performance as the chaos shift keying (CSK) system.