Approximate optimal demodulation for multi-user binary coherent chaos-shift-keying communication systems

A maximum likelihood (ML)-based approximate optimal demodulator is derived for multi-user coherent chaos-shift-keying (CSK) communication systems. Starting from the ML approach, the theoretically optimal demodulator is derived first. Then, the optimal demodulator is approximated by considering the dynamical property underlying two successive values in the transmitted chaotic sequence. A detailed example is given to show how this approximate optimal demodulator can be applied in the case of Bernoulli shift map spreading. The bit error rate (BER) performance is simulated for comparison with the common correlation demodulator. The paper's main contribution is to show that the performance of a multi-user CSK communication system can be improved by optimal demodulation in the sense that it then out-performs the BER lower bound from correlation demodulation.