Development of New Filter and Tracking Schemes for Weak GPS Signal Tracking

Various emerging applications require location of users in challenging environments where typical GPS receivers suffer degraded performance or complete failure. Special algorithms and techniques are required to track weak GPS signals, where the signal is typically weaker by 10 to 40 dB compared to the nominal or line-of-sight signal strength. This thesis endeavours to propose solutions that can potentially offer performance improvements over conventional techniques.;For the cases when external data aiding is not available, a decision feedback principle is used herein, in which the data bits are estimated through the tracking process itself. An enhanced digital phase locked loop with a frequency rate estimator is also developed. The NCO with phase rate and frequency rate feedback is introduced and based on this NCO and the transfer function of the frequency rate estimator, the tracking loop is optimized in order to minimize the phase noise variance. By utilizing this loop, the performance of low update rate loops in terms of phase mismatch and bit error rate can be improved. A multistage tracking scheme is also implemented to overcome the problem of tracking weak GPS signals in indoor environments. In this technique several tracking schemes are serially cascaded. It is shown that this technique combined with a developed optimum delay locked loop can be used for seamless outdoor to indoor tracking.;The performance and advantages of these techniques are shown based on a developed software GPS signal simulator, hardware simulated signals, live attenuated GPS signals and live GPS signals in selected indoor environments. The tracking capabilities of these schemes are also compared with a commercially available high sensitivity GPS receiver.;Optimum digital tracking filters for loops of first to fourth order, for rate only and phase and rate feedback NCO are derived. It is shown that, contrary to conventional methods, the loops remain stable for high B LT (the product of loop noise bandwidth and loop update interval) values and for both types of aforementioned NCOs. By using these filters, a significant improvement for high BLT can be achieved, allowing one to operate in ranges where previous methods cannot operate. As a result, stable loops with longer integration times (update interval) can be easily designed and the tracking sensitivity is improved accordingly.