Space-time blind equalisation in impulsive noise

The presence of non-Gaussian impulsive noise in wireless system can degrade the performance of existing equalisers and signal detectors. The problem of blind source separation and equalisation for multiple-input/multiple-output (MIMO) channels under heavy-tailed impulsive noise environment is studied. A generalised multi-user constant modulus (CM) cost function is proposed by employing the fractional lower order CM property of the equaliser input signals as well as the fractional lower order cross-correlations between them. The associated adaptive blind equalisation algorithm based on a stochastic gradient descent method is defined as fractional lower order multi-user constant modulus algorithm (FMU/CMA), which is able to mitigate impulsive channel noise while recovering all input signals simultaneously. The steady-state mean-square error (MSE) performance of the FMU/CMA is studied in a noise-free environment; the approximate expression is derived based on the energy-preserving relation and the Taylor series expansion. Simulation studies are undertaken to support the analysis with CM and non-CM signal.