Space-Time Blind Equalization of Dispersive MIMO Systems Driven by QAM Signals

Jin Li,Da-Zheng Feng,Bingbing Li

DOI: https://doi.org/10.1109/tvt.2018.2790432

IF: 6.8

2018-01-01

IEEE Transactions on Vehicular Technology

Abstract:This paper investigates space-time blind equalization (BE) of dispersive multiple-input multiple-output communication systems driven by high throughput quadrature amplitude modulation signals. Multistage processing is adopted to overcome the severe local convergence problem in BE. First, a good initial value for the equalizer is obtained by forcing one of the system inputs with desired delay to be recovered. Then, the fast multimodulus algorithm (MMA) is used to roughly equalize the communication system. After that, an improved MMA (IMMA) is developed to effectively search for a fine equalizer for the system. Moreover, the novel modified Newton method (MNM) proposed in our previous work is employed to fast optimize the MMA and IMMA cost functions, which significantly reduces the computational load. Furthermore, the proposed algorithm has an additional benefit that once a signal is recovered, its corresponding single-input multiple-output channel impulse response is estimated by the classical least squares methods. Then, the influence of the recovered signal to the original received signals is removed to avoid extracting this recovered signal repeatedly. The quadratic rate of convergence of the MNM is theoretically analyzed, and the good equalization performance of the IMMA is explained. Finally, simulation results are provided to illustrate the effectiveness of the proposed algorithm.