A time-channelized programmable canonical correlation analyzer

In many communication signal environments, the presence of cochannel interference may result in significant system performance degradation. In these environments, it is necessary to mitigate against such interference in order to extract the desired signals of interest. One approach to this problem is to utilize processing techniques that exploit known signal properties or characteristics to discriminate between signals of interest and undesired interference. The time-channelized programmable canonical correlation analyzer (TC-PCCA) is proposed as a single-sensor technique for separating cochannel signals by exploitation of temporal redundancy and known signal properties. The proposed technique temporally partitions a received signal block into subblocks exhibiting inter-subblock redundancy (e.g., redundancy due to repeated data or training sequences). The subblocks are simultaneously processed by the PCCA to extract the desired signals in a manner similar to blind adaptive spatial filtering. Specific realizations of the TC-PCCA are considered for several signal types, including the digital forward control channel signal employed in the AMPS cellular system. Signal extraction and synchronization performance are characterized by analysis and simulation.