An adaptive phase processing technique applied to the recovery of a coherent reference

The adaptive numerical phase processor (ADNUPP) is discussed, which applies a simple implementation of a minimum variance estimator (Kalman filter) to obtain an estimate of carrier phase and frequency coherence. This technique allows the acquisition of the carrier phase in a deterministic time frame with a minimum of phase jitter as described by the estimation and measurement noise parameters in the applied algorithm. This implementation allows the bandwidth of the recovery system to vary continuously based on the confidence of the estimate, as opposed to the discrete mechanisms typically used. As a result of this implementation, wideband operation can be initially maintained and convergence on an estimate is achieved in the steady-state condition. A system was selected for comparing the performance of receiver architectures using a standard implementation (PLL) and one utilizing the ADNUPP. A pseudo-random BPSK modulation waveform was generated and degraded with additive white Gaussian noise (AWGN). The signal was then filtered and demodulated. A comparison of the acquisition time (normalized to the data rate) for the ADNUPP and traditional PLL designs is plotted.<
>