A Robust Periodically Time Varying (PTV) Filtering System for Real-world Collection Environments

An SBIR study is proposed for a robust and versatile periodically time varying (PTV) filtering workstation for separation of spectrally overlapping wide-sense cyclostationary (WSCS) signals, and to implement and demonstrate key workstation components on realistic computer-simulated signals and environments. This study emphasizes stressful conditions that typify real-world signal collection missions, and the need to separate overlapped signals with little or no a priori knowledge of their modulation types, carrier frequencies, baud rates, or other parameters. The key component of the signal separation workstation is the biresolution PTV (BRPTV) filter, a blind, adaptive PTV filter that overcomes the interference capture problem that impedes use of conventional blind PTV and non-PTV adaptive filters in many applications. The BRPTV filter is rapidly convergent and includes a variety of known modulus and known constellation cost functions. A specialized adaptive PTV filter termed the optimum linear CDMA signals such as in frequency re-use cellular radio networks. The OLCR is highly computationally efficient and exploits the orthogonality of the CDMA code set more fully than conventional CDMA receivers, to perform code-nulling on interfering CDMA signals. The OLCR can separate larger numbers of CDMA signals than conventional approaches, achieves better separation performance, and is robust over larger signal power disparities. An option is included to implement the decision-theoretic cyclic spectrum analyzer (DTCSA) for initial detection and characterization of overlapped signals. The DTCSA overcomes limitations of conventional CSAs, including large computational requirements, slow convergence, dynamic range limitations, large-signal artifacts, tone-induced clutter, and high CSA surface display density. The DTCSA provides signal parameters to the BRPTV filter and synergistically coordinates with it to provide a powerful and flexible signal separation capability.