Electronic Version

The article provides an overview of the BeiDounavigation message contents and highlights its specificcommunalities and differences with respect to otherGNSS constellations. Making use of data collected bymulti-GNSS monitoring stations of the MGEX andCONGO networks, the quality of BeiDou broadcastephemerides is assessed through the analysis of satellitelaser ranging measurements, comparison with postprocessed orbit and clock products as well as positioningtests. Specific attention is given to signal-specific groupdelays and their proper consideration in the positioning.

A Bayesian method for dynamical offline estimation ofthe position and the path loss model parameters of awireless network's communication node is presented.Two versions of three different online positioningmethods are tested using data collected from cellularnetworks and WLAN networks in outdoor and fromWLAN networks in indoor environments. The tests showthat the methods that use the estimated path lossparameter distributions with finite precisions outperformthe methods that only use point estimates for the pathloss parameters. They also outperform the coverage areabased positioning method and path loss model methodwith generic path loss parameters, and are comparable inaccuracy with the k-nearest neighbour fingerprintingmethod. Taking the uncertainties into account iscomputationally demanding, but the Gauss–Newtonoptimization methods is shown to provide a goodapproximation with computational load that isreasonable for many real-time solutions.

An advanced Receiver Autonomous IntegrityMonitoring (ARAIM) approach is investigated whenaugmenting GPS satellites with the current regionalBeiDou constellation. A procedure for integritymonitoring, including checking its availability, faultdetection and exclusion, and integrity testing ispresented. Fault modes and their probabilities using GPSand GPS+BeiDou are discussed. Testing of ARAIM forvertical guidance using real data in eight sites distributedglobally (Australia, China, Netherlands, eastern Canadaand Peru) show that the addition of the BeiDouconstellation, despite the decreased preliminaryconfidence placed in its performance compared withGPS, results in a substantial improvement to ARAIMavailability performance and a higher level of integrity,in particular at sites observing all of its currentconstellation (Australia and China). The improvementwas less in sites that can only observe some or no GEOand IGSO satellites (Netherlands, Canada and Peru).However, the benefit of adding BeiDou to GPS at thesesites is expected to substantially improve with fulldeployment of MEO satellites.

In this paper, an integrated inter-vehicles wirelesscommunications and positioning system supportingalternate positioning techniques is proposed to meet therequirements of safety applications of CooperativeIntelligent Transportation Systems (C-ITS). Recentadvances have repeatedly demonstrated that road safetyproblems can be to a large extent addressed via a rangeof technologies including wireless communications andpositioning in vehicular environments. The novelcommunication stack utilizing a dedicated frequencyspectrum (e.g. at 5.9 GHz band), known as DedicatedShort-Range Communications (DSRC), has beenparticularly designed for Wireless Access in VehicularEnvironments (WAVE) to support safety applications inhighly dynamic environments. Global NavigationSatellite Systems (GNSS) is another essential enabler tosupport safety on rail and roads. Although currentvehicle navigation systems such as single frequencyGlobal Positioning System (GPS) receivers can provideroute guidance with 5-10 meters (road-level) positionaccuracy, positioning systems utilized in C-ITS mustprovide position solutions with lane-level and even inlane-level accuracies based on the requirements of safetyapplications. This article reviews the issues and technicalapproaches that are involved in designing a vehicularsafety communications and positioning architecture; italso provides technological solutions to further improvevehicular safety by integrating the DSRC and GNSSbased positioning technologies.

In Global Navigation Satellite System (GNSS)positioning, ranging signals are delayed when travellingthrough the ionosphere, the layer of the atmosphereranging in altitude from about 50 to 1000 km consistinglargely of ionized particles. This delay can vary from 1meter to over 100 meters, and is still one of the mostsignificant error sources in GNSS positioning. In preciseGNSS positioning applications, ionospheric errors mustbe accounted for. One way to do so is to treat unknownionosphere delay as stochastic parameter, which canaccount for the ionospheric errors in the GNSSmeasurements as well as keeping the full originalinformation. The idea is adding ionospheric delay fromexternal sources as pseudo-observables. In this paper, theperformance of ionosphere-weighted model is evaluatedusing real data sets, and the correctness of prioriionosphere variance is also validated.

Dynamic orbit determination is the conventionaltechnique that has been commonly used for precise orbitdetermination (POD) of satellites at various orbitalaltitudes. The performance of this technique is mainlylimited by inaccurate modelling of force perturbationsacting on satellites. The perturbations include the Earth’sgravity field, atmospheric drag, solar radiation pressureetc. Due to the fact that low Earth orbit (LEO) satellitesare highly sensitive to the Earth’s gravity field, theaccuracy of the gravity field model used in the dynamicPOD technique directly affects the accuracy of POD ofLEO satellites. Therefore, selection of an accurategravity field model for improving the POD accuracyplays a significant role in meeting the stringentrequirements of space applications such as radiooccultation, remote sensing and altimetry. Nowadays,with the successful launches of the CHAMP,GRACEand GOCE gravity missions, various high accuracygravity field models have been developed and madepublicly available at the International Centre for GlobalEarth Models (ICGEM).In this study, the performance of13 selected gravitymodels applied in the dynamic POD was assessed usingspace-borne dual-frequency GPS measurements from thetwin GRACE satellites during the period from 1st to 31stMarch 2008, and the effects of time-varying low-degreespherical-harmonic coefficients C20 , C30 and C40 onPOD for the twin GRACE satellites were also analysed.The results of tracking data residuals, orbital overlap,external orbit comparison and independent satellite laserranging (SLR) validation demonstrated that the highestPOD accuracies of GRACE-A and -B are about 2.1 cmand 2.7 cm with respect to SLR measurementsrespectively and this is achieved using those combinedmodels ,i.e. EIGEN-51C, GO_CONS_GCF _2_DIR_R3,andGOCO03S.In addition, a comparison of the orbitsgenerated with and without the time-varying gravityfield indicates that orbit variability caused by the timevarying component of EIGEN-GL04S1 was at a few mmlevel, suggesting that the time-varying low-degreespherical-harmonic coefficients do not lead to a notablevariability in orbit quality

Stochastic modelling is critical in GNSS data processing.Currently, GNSS data processing commonly relies onthe empirical stochastic model which may not reflect theactual data quality or noise characteristics. This paperexamines the real-time GNSS observation noiseestimation methods enabling to determine theobservation variance from single receiver data stream.The methods involve three steps: forming linearcombination, handling the ionosphere and ambiguitybias and variance estimation. Two distinguished waysare applied to overcome the ionosphere and ambiguitybiases, known as the time differenced method andpolynomial prediction method respectively. The realtime variance estimation methods are compared with thezero-baseline and short-baseline methods. The proposedmethod only requires single receiver observation, thusapplicable to both differenced and un-differenced dataprocessing modes. However, the methods may be subjectto the normal ionosphere conditions and lowautocorrelation GNSS receivers. Experimental resultsalso indicate the proposed method can result on morerealistic parameter precision.