Breaking a chaotic direct sequence spread spectrum communication system using interacting multiple model-unscented Kalman filter

Abstract

In this paper, a new method to break chaotic direct sequence spread spectrum (CD3S) communication systems is proposed. Here, the CD3S communication system transmitting different information symbols is considered as a combination of two subsystems which are driven by two different chaotic dynamic models, respectively. At every single time moment, the CD3S signal can be regarded as generated by the subsystem corresponding to the information symbol transmitted. Then, based on the multiple model form of CD3S signals, an interacting multiple model unscented Kalman filter with model switching detection mechanism is exploited to track the CD3S signals. The l(2)-norm of tracking errors is used to choose the model which best matches the intercepted signals. Thus, the information symbols are recovered indirectly. Compared with the existing methods, the proposed algorithm can: (1) reduce the influence of a low spreading factor; (2) calculate the spreading factor using the length of time intervals between model switching; and (3) be more effective under scenarios of low signal-to-noise ratio or multipath fading. Simulation results verify the superiority of the proposed method.