Mean time to loose lock of phase tracking by particle filtering

We consider the classical problem of phase lock on frequency modulation (FM) signals using the recent methods of particle filters (PF). This problem is nonlinear in nature, and building an optimal filter is impossible. State-of-the-art suboptimal phase estimators are based on phase lock loops (PLL) which are equivalent to an extended Kalman filter (EKF) realization. We show that applying PF methods, which preserve the nonlinearity of the phase measurements, result in improved performance of 0.94 dB carrier-to-noise (CNR) ratio considering the mean time to lose lock (MTLL). We show simulation results for the PLL-EKF and the PF estimators, compare their performance and demonstrate how an increase in the number of particles improves the performance of the particle filter estimator.