Design of cosine modulated filterbank transmultiplexer for unknown channels using genetic algorithm

Filterbank transmultiplexer (FB TMUX) system has been considered as a kind of multicarrier communication technology. When there are interferences, the FB TMUX system has better performance than orthogonal frequency division multiplexing (OFDM) system, as filters which are used in the FB TMUX have better frequency responses than those of the OFDM. The FB TMUX can be efficiently implemented by using cosine modulated filter banks (CMFB). As the CMFB TMUX is obtained from lowpass filters, design of the lowpass filters determines the performance of the CMFB TMUX. Recently the design of CMFB TMUX for unknown channels as been introduced. But in this design roll off factors of the lowpass filters and channel delay compensation factor are not optimized. In this work genetic algorithm (GA) is used for optimizing the roll off factors of the lowpass filters and channel delay compensation factor to improve the signal to interference ratio (SIR) of the CMFB TMUX. Simulation results show that the SIR is considerably improved by using the optimum values of these parameters.