Single-Channel Blind Identification with FIR Filter Design in Frequency Domain

Abstract

The problem of estimating the communication channel affecting a signal, falls into the category of blind system identification. When only one version of the signal is available, this is particularly difficult problem. This paper presents a simple single-channel blind identification of channel impulse response which can be low-pass, band-pass or highpass filter, etc with a FIR filter design in frequency domain. The proposed algorithm is tested with different type of input sources (such as random sequence and speech) using different kind of channels. It was shown that, the proposed algorithm sounds to be more effective for random sequences and can also be used in different application such as loudspeaker or telephone channel equalization.