Subspace blind MIMO channel equalization with quadratic complexity

An SOS-based blind equalization algorithm for the SIMO channel has recently been proposed that has an unprecedented quadratic complexity in the channel memory, compared to all existing SOS-based techniques whose complexity is cubic in this regard. In this paper, we show that this technique can be adapted to the MIMO channel and we prove that its complexity is maintained independently from the number of input/output channels. Simulation tests are reported that sustain the feasibility of this technique in practical observation conditions.