Study On High Dynamic GPS Carrier Tracking Techniques

Along with GPS navigation and missile-guiding systems used in modern wars more and more frequently, research on the applications of GPS system in high dynamic circumstance brings increasing attentions. In high dynamic circumstance, there exist very high Doppler frequency shift and its derivatives on the carrier received by GPS receivers, therefore, conventional GPS carrier tracking loops cannot secure the normal works. Aiming at problem of precisely carrier tracking on the condition of high dynamic and low signal-to-noise ratio, thorough and systematic studies carry out in this dissertation, and some benefit results acquired.Firstly, a DDS-based method for digital-domain synthesis of time-varying Doppler frequency in dynamic GPS signal is proposed in this dissertation. The scheme can efficiently design the frequency controlling words of carrier NCO in the synthesizer so as to accurately simulate received carrier Doppler frequency shift and its derivatives. Therefore, the scheme can supply necessary prior data for the study of high dynamic GPS carrier tracking techniques.Aiming at problem of receiving GPS signals in high dynamic environments, detailed analysis in theory is performed in this dissertation, and the performances of conventional GPS carrier tracking loops on the condition of high dynamic and low signal-to-noise ratio are obtained by means of numerical simulations. Simulation results indicate that in typically high dynamic environment, loss-of-lock thresholds of conventional PLL and CPAFC are 25.5dB-Hz and 24.5dB-Hz respectively, especially when input CNR of the loops is typically low as 25dB-Hz, RMS tracking errors of them are respectively 60Hz and 28.6Hz, which clearly shows their limits in high dynamic environments. Aiming at problem of precisely carrier tracking on the condition of high dynamic and low signal-to-noise ratio, four schemes are proposed in this dissertation.A new structure of automatic frequency controlling loop based on overlapping DFT algorithm is proposed in this dissertation. The scheme combines thoughts of classical power spectrum estimation and basic theory of conventional DPLL, and its discriminator can be thought of as an extension of that in CPAFC when the number of sampling points once treated in the discriminator is augmented. Therefore, in theory, tracking precision of the proposed scheme will precede CPAFC. Simulation results indicate that in typically high dynamic environment, loss-of-lock threshold of the scheme is 24dB-Hz, especially when the input CNR is typically low as 25dB-Hz, RMS tracking error of the scheme is 7Hz. Since phase or frequency discriminators and loop filters used in conventional PLL-based carrier tracking scheme are not designed conforming to optimal criteria, performance of conventional loop is very limited. Aiming at this problem, a new quasi-optimal structure of GPS carrier tracking loop based on extended Kalman filtering is proposed in this dissertation. As the scheme is designed conforming to the criteria of approximate linear least-mean-square error estimation, it is theoretically quasi-optimal. Simulation results indicate that in typically high dynamic environment, loss-of-lock threshold of the scheme is 24.5dB-Hz, especially when the input CNR is typically low as 25dB-Hz, RMS tracking error of the scheme is 2.4Hz.Aiming at problem that estimation precision of the extended Kalman filter deteriorates due to linearization, another new quasi-optimal structure of GPS carrier tracking loop based on unscented Kalman filtering is proposed in this dissertation. Since the unscented Kalman filtering-based loop can exploit information beyond 2nd order moments about distribution during the course of estimation, theoretically, it over performs the extended Kalman filtering-based loop. Simulation results indicate that in typically high dynamic environment, loss-of-lock threshold of the scheme is 24.5dB-Hz, especially when the input CNR is typically low as 25dB-Hz, RMS tracking error of the scheme is 2.2Hz.Finally, a theoretically optimal structure of GPS carrier tracking loop based on maximum likelihood estimation and linear Kalman filtering is proposed in this dissertation. On the condition that prior-distribution of estimated parameter is unknown, estimate supplied by maximum likelihood estimation is provided with least mean square error; on the other hand, as for states filtering for linear system, Kalman filtering is optimal in theory. Therefore, the scheme which incorporates maximum likelihood estimation and linear Kalman filtering ought to be the optimal carrier tracking scheme in theory, and it will over performs the above two quasi-optimal schemes. Simulation results indicate that in typically high dynamic environment, loss-of-lock threshold of the scheme is only 23dB-Hz, especially when the input CNR is typically low as 25dB-Hz, RMS tracking error of the scheme is only 1.1Hz.