Novel VLSI architecture for real-time blind source separation

Blind source separation of independent sources from their convolutive mixtures is a problem in many real-world multi-sensor applications. In this paper, we propose efficient hardware architecture for the realization of real-time BSS and implement it in Verilog HDL. The design implements the modified Torkkola´s BSS algorithm for audio signals based on Infomax technique. We use feedback network architecture with non-causal filters instead of the causal filters. Description of the hardware, as well as discussion of some issues regarding Verilog implementation, along with the results of simulation are presented.