Blind Equalization via Approximate Maximum Likelihood Source Separation

Blind equalization of single input multiple output (SIMO) FIR channels can be reformulated as the problem of blind source separation. It was shown in [4] that the natural gradient-based source separation method could recover source successfully for ill-conditioned channels since it has the equivariant property. However, the e ect of additive noise was not considered in [4]. In this letter we develop a new approximate maximum likelihood source separation (AMLSS) method using the natural gradient and apply it to the task of blind equalization. We show that the proposed method outperforms the BSBE [4] in the presence of Gaussian noise. Indexing terms: Blind equalization, blind source separation, maximum likelihood, natural gradient. Electronics Letters, vol. 37, no. 1, pp. 61-62, January 27 2001 Please address correspondence to Prof. Seungjin CHOI, Department of Electrical Engineering, Chungbuk National University, 48 Kaeshin-dong, Cheongju, Chungbuk 361-763, KOREA, Tel: +82-43-261-2421, Fax: +8243-263-2419, Email: [email protected]