Fast Time to Fine Time Method to Improve First Fix Accuracy with Modernized Signals in Urban Canyons

Abstract

Better than one meter measurement precision of linearized GNSS pseudoranges requires fine-time: defined as a system time error less than one millisecond. Without fine-time, the linearized pseudorange measurement incurs a computation error equivalent to radial component of the satellite motion between the true and estimated transmission time. This error is referred to as fictitious measurement error and it is additive to the thermal, atmospheric, and multipath errors imposed the code-phase measurements. This condition occurs commonly in mass market receivers that produce a rapid first position fix with a coarse-time solver before a GNSS receiver is able to decode satellite time from the navigation message. This paper presents a method of fusing the coarse-time solver with additional time information available in the physical layer of modernized signals: namely the difference between the measured and predicted secondary code-phase. These time sources provide fine-time but have an ambiguity equal to the length of the secondary code. The method identifies the most likely rounded time estimates among a set of candidate times as the solution with the lowest posterior residuals. With long secondary codes, the fictitious measurement error will dominate at the wrong candidates. Large measurement errors prevent identifying a clear minimum in the posterior residuals across the candidate solutions. An outlier detection and mitigation method are required to remove the larger measurement errors.