Multi-Constellation GNSS Multipath Mitigation Using Consistency Checking

In a typical urban environment, a mixture of multipath-free, multipath-contaminated and non-line-of-sight (NLOS) propagated GNSS signals are received. The errors caused by multipath-contaminated and NLOS reception are the dominant source of reduced consumer-grade positioning accuracy in the urban environment. Many conventional receiver-based and antenna-based techniques have been developed to mitigate either multipath or NLOS reception with mixed success. Nevertheless, the positioning accuracy can be maximised based on the simple principle of selecting only those signals least contaminated by multipath and NLOS propagation to form the navigation solution. The advent of multi-constellation GNSS provides the opportunity to realise this technique that is potentially low-cost and effective for consumer-grade devices. It may also be implemented as an augmentation to other multipath mitigation techniques. The focus of this paper is signal selection by consistency checking, whereby measurements from different satellites are compared with each other to identify the NLOS and most multipath-contaminated signals. The principle of consistency checking is that multipath-contaminated and NLOS measurements produce a less consistent navigation solution than multipath-free measurements. RAIM-based fault detection operates on the same principle. Three consistency-checking schemes based on single-epoch least-squares residuals are assessed: single sweep, recursive checking and a hybrid version of the first two. Two types of weighting schemes are also considered: satellite elevation-based and signal C/N-0-based weighting. The paper also discussed the different observables that may be used by a consistency-checking algorithm for different applications and their effect on detection sensitivity. Test results for the proposed algorithms are presented using data from both static positioning and stand-alone dynamic positioning experiments. The static data was collected using a pair of survey-grade multi-constellation GNSS receivers using both GPS and GLONASS signals at open sky and urban canyon locations, while the dynamic data was collected using a consumer-grade GPS/GLONASS receiver on a car in a mixed urban environment. Significant improvements in position domain are demonstrated using the weighted recursive methods in the open environments. However in the urban environments, there are insufficient directly received signals for the conventional RAIM-based signal selection to be effective all the time. Both positioning improvements and risky outliers are demonstrated. More advanced techniques have been identified for investigation in future research.