Innovations and singular value decomposition for blind sequence detection in wireless channels

= ' blind sequence detection (BSD) aigorithm based on the innovations approach is proposed and its performance in a Rayleigh fading environment is evaluated. A cornparison between the innovations and Tong's Singular Value Decomposition (SVD) [25] based blind sequence detection algorithm is also presented. Further insight is gained on how blind sequence detectors behave by examining several parameters such as differential encoding, signal subspace dimension d , correlation delay k and diversity. When diversity is no t used, the innovations approach t O blind sequence detection ciearly outperforms the SVD based blind sequence detector . Overail t hough. the SVD based blind sequence detec tion algorît hm has a bet t er performance than the innovations approach. However, if one looks a t the 'combined performance' and takes into account that the innovations approach is simple, causal, recursive and numericaiiy efficient when compared to SVD , then the performances of the two blind sequence detectors are comparable.