Digital filtering and prediction for communications systems time synchronization

A timing subsystem consisting of an
N
-point finite impulse response (FIR) digital differentiator followed by an
M
-step adaptive linear predictor is described. The techniques may be employed for initial timing acquisition and subsequent resynchronization of ASK, FSK, and PSK digital communications systems employing pilot signals. The overall approach is effective, and provides a new method for solving a most critical problem ultimately affecting the achievable probability of bit-error rate. Various tradeoffs are discussed for selecting the type of digital differentiator, as well as the parameter of both the differentiator and predictor. Performance data is presented for realistic signal-to-noise ratios and digital word sizes. The described method has been implemented as part of a microprocessor-based FSK acoustic telemetry system, and is found to be computationally efficient as well as extremely accurate.