|
UWEETR-2004-0003 Author(s): Keywords: Abstract We evaluate by simulation the performance of a decision-directed adaptation technique for the passive phase conjugation (PPC) array equalizer, a coherent receiver structure for underwater acoustic communications channels. Empirically-derived models recreate the acoustic channels of three at-sea telemetry experiments, providing telemetry simulations of many data packets over various link configurations. From these simulations we collect performance statistics of the decision-directed PPC equalizer. The algorithm, which we call DDPPC, uses past symbol detections to build linear statistical models for the channel impulse responses (CIRs) in a block fashion. Using the LSQR iterative solver for non-square systems combined with a special initialization method allows efficient and compliant channel estimation and tracking. PPC then uses the CIR estimates directly in its equalization technique, without solving a secondary set of equations for equalizer coefficients. The telemetry experiments generating the simulator models involved ranges of 0.9 to 4.6 \km\ under windy surface conditions. We execute statistical trials over multiple data packet realizations under BPSK and QPSK modulations at different bandwidths, signaling at 2174 symbols/s. Array geometry is varied between sparse vertical receiver apertures of two to 14 elements, with the largest spanning the water column |