Robust receiver for OFDM-DCSK modulation via rank-1 modeling and ℓp-minimization

Abstract In this paper, the problem of receiver design for orthogonal frequency division multiplexing differential chaos shift keying (OFDM-DCSK) communication systems is addressed. By exploiting the rank-1 property of the symbol matrix, we propose to apply dimensionality reduction on the time-domain data symbols received from the OFDM-DCSK transmitter for noise reduction, followed by chaotic demodulation on the resultant symbols to decode the information bits. In the presence of additive white Gaussian noise (AWGN), the rank-1 matrix approximation can be simply achieved by the truncated singular value decomposition, corresponding to the solution of l 2 -norm minimization. While for impulsive noise environments such as in power line communication systems, we develop an alternating optimization algorithm for l p -based matrix factorization, where 0 p 2 . The bit error rate (BER) of our approach in AWGN is also analyzed and verified. Simulation results demonstrate that the devised receiver is superior to the conventional OFDM-DCSK method in terms of BER and root mean square error performance for AWGN as well as impulsive noise including the Middleton class A distribution and α -stable process.