부산대학교 도서관

A recursive linear detection algorithm is proposed for the detection of signals from an asynchronous direct-sequence code-division multiple-access (DS-CDMA) communication system. This algorithm works for short as well as long codes. Under some reasonable conditions, this algorithm is proved to be stable and converges to the ideal decorrelating detector (IDD) with a sufficiently large memory length. The performance of the algorithm is analyzed in some detail. Upper and lower bounds for the bit-error probabilities are developed. It is demonstrated that the two bounds converge to the bit-error probabilities of the IDD as the large memory length increases. Simulation results show that the recursive detector proposed in this paper outperforms the truncated decorrelating detector with less memory and less computational complexity. Index Terms--Banach space, bounded operator, decorrelation, Hilbert space, least squares estimation (LSE), recursive detection.