Blind equalisation algorithm of FIR MIMO system in frequency domain
Xin Shen, Hao-Ke Xu, Fengyu Cong, J. Lei, Kun Huang, Y. Zhang, Guang Meng · IEE Proceedings - Vision Image and Signal Processing · 2006
Blind equalisation of finite impulse response multi-input multi-output systems is an important task for numerous applications such as speech separation, dereverberation, communication, signal processing and control. A cost function is constructed based on second-order statistics of the equaliser output, and by minimising it, two algorithms are derived, respectively, with conventional and natural gradient. The algorithms are proved to adapt to non-stationary signals. In the algorithms, the learning rate is multiplied by a constraint factor, which is adjusted adaptively. Comparison and simulation indicate the ability of the algorithms to perform blind equalisation under the weaker condition (the finite impulse response system is equalisable).