Filterbank-based blind code synchronization for DS-CDMA systems in multipath fading channels

We present a filterbank approach to blind code synchronization for asynchronous direct-sequence (DS) code-division multiple-access (CDMA) systems. The key idea of the proposed scheme is to first pass the received signal through a bank of filters, which are designed to enhance signals of interest and suppress interference/noise, and then to derive the code timing from the filtered data. The only required knowledge by the proposed filterbank scheme is the spreading code of the desired user. It can be used in various environments, including frequency-nonselective and frequency-selective, time-invariant, and time-varying fading channels. It can deal with colored channel noise and unmodeled interference, such as inter-cell interference (ICI) and narrowband interference. It has relatively low complexity and can be readily implemented using standard adaptive algorithms. We show that under mild conditions, the proposed scheme yields statistically consistent [in signal-to-noise ratio (SNR)] code timing estimates, irrespective of the strength of the interference and with only a finite number of data samples. We also derive an unconditional Cramer-Rao bound (UCRB), which serves as a lower bound for all unbiased blind code synchronization schemes. Numerical results indicate that the proposed scheme compares favorably with a popular subspace-based method in terms of user capacity, near-far resistance, and robustness to time-varying fading and unmodeled interference.