Blind adaptive equalizer for broadband MIMO time reversal STBC based on PDF fitting

This work presents a new blind multiuser equalization strategy for Time Reversal Space Time Block Coding (TRSTBC) signals transmitted over a dispersive MIMO channel. The adaptation is based on forcing the probability density function (PDF) of the equalizer output to match the desired PDF of corresponding source symbols. In the proposed approach, the PDFs are estimated by means of the Parzen window method using Gaussian Kernels. The cost function combines this PDF fitting with an orthogonality criterion derived from the TR-STBC structure. The performance is demonstrated in a number of simulations and benchmarked against other blind schemes. The proposed algorithm has a moderate computational complexity and can perform with higher adaptation rate.