The Effect of Modeling Clock Dynamics on Kalman Filter Based GNSS Signal Tracking

Tracking loops are the most popular scheme to carry out carrier phase and frequency estimation in real-time Global Navigation Satellite System (GNSS) receivers. This is in part because of their very low computational complexity, but new architectures have emerged as a consequence of increasing capability of embedded systems to perform high-cost computing. This enables the implementation of tracking schemes that make the modeling of new error sources easier. This work presents a Kalman Filter based tracking architecture and the modeling of clock dynamics in the formulation. The most common way to characterize clock dynamics is modeling the Allan variance of the oscillators, on which this work focuses. The importance of clock dynamics modeling in GNSS signal tracking is analyzed through simulations and results show that tracking sensitivity and estimation error increase in the presence of clock errors. The simplicity of adding new error models to a Kalman Filter formulation is showed. The dynamic model of the filter is designed considering the clock dynamics and a simulation performance is presented in a high dynamics scenario such as a space vehicle launch. The same algorithm was run feeding it with samples obtained from a GNSS receiver front-end to analyze its behavior under real clock effects. with good results.