Design and analysis of a generalized DLL/FLL discriminator for GPS receivers

High-performance discriminators are of a significant interest in GPS receiver tracking-loop design. Main lobes of both the auto-correlation function of the band-limited pseudorandom noise code and the sinc function resulting from the carrier frequency tracking, resemble the shape of an inverted parabola. This interesting observation led to the proposal of a generalized DLL/FLL discriminator based on parabola fitting, which is the topic of this research. To provide better insight into the mechanism of this discriminator, a semi-analytical noise performance prediction method is presented based on the Gauss–Hermite cubature transformation, and simple loss functions are defined to comprehensively measure the S-curve-related performances. On this basis, design parameter optimization for the generalized DLL/FLL discriminator is addressed by trade-off analysis. Controlled experiments with real and simulated signals show that the overall performance of the generalized discriminator is significantly better than the commonly used DLL and FLL discriminators and is more suitable for use in applications that need to handle weak signal, high dynamics and/or large offsets in frequency and code phase estimation during the acquisition to tracking transition.