Navigation Recommender:Real-Time iGNSS QoS Prediction for Navigation Services

摘要

Global Navigation Satellite Systems (GNSSs), especially Global Positioning System (GPS), have become commonplace in mobile devices and are the most preferred geo-positioning sensors for many location-based applications. Besides GPS, other GNSSs under development or deployment are GLONASS, Galileo, and Compass. These four GNSSs are planned to be integrated in the near future. It is anticipated that integrated GNSSs (iGNSSs) will improve the overall satellite-based geo-positioning performance. However, one major shortcoming of any GNSS and iGNSSs is Quality of Service (QoS) degradation due to signal blockage and attenuation by the surrounding environments, particularly in obstructed areas. GNSS QoS uncertainty is the root cause of positioning ambiguity, poor localization performance, application freeze, and incorrect guidance in navigation applications. udIn this research, a methodology, called iGNSS QoS prediction, that can provide GNSS QoS on desired and prospective routes is developed. Six iGNSS QoS parameters suitable for navigation are defined: visibility, availability, accuracy, continuity, reliability, and flexibility. The iGNSS QoS prediction methodology, which includes a set of algorithms, encompasses four modules: segment sampling, point-based iGNSS QoS prediction, tracking-based iGNSS QoS prediction, and iGNSS QoS segmentation. Given that iGNSS QoS prediction is data- and compute-intensive and navigation applications require real-time solutions, an efficient satellite selection algorithm is developed and distributed computing platforms, mainly grids and clouds, for achieving real-time performance are explored. The proposed methodology is unique in several respects: it specifically addresses the iGNSS positioning requirements of navigation systems/services; it provides a new means for route choices and routing in navigation systems/services; it is suitable for different modes of travel such as driving and walking; it takes high-resolution 3D data into account for GNSS positioning; and it is based on efficient algorithms and can utilize high-performance and scalable computing platforms such as grids and clouds to provide real-time solutions.udA number of experiments were conducted to evaluate the developed methodology and the algorithms using real field test data (GPS coordinates). The experimental results show that the methodology can predict iGNSS QoS in various areas, especially in problematic areas.ud