Parameter Estimation and Prediction of the Chirp and Stochastic Pulse Position Modulation Combined Signal

Recent work has proposed a certainty trend (CT) elimination technique employed for the auto-regressive/autoregressive and moving-average (AR/ARMA) model pulse position prediction. In this paper, we investigate the intra pulse parameter estimation and pulse position prediction of the chirp and stochastic pulse position modulation (CSPPM) combined signal. The quick dechirp method is adopted to the initial frequency and chirp rate estimation. To get a stationary data series satisfying the premise condition of the AR/ARMA model prediction, a least square fitting (LSF) scheme to remove the CT term contained in pulse position sequence is presented. Compared with the classic logarithmic difference conversion (LDC) smooth method, AR/ARMA prediction performance via LSF has a significant improvement, about 70% to 96% for AR prediction and 42% to 99% for ARMA prediction.