GPS World, March 2011

TRANSPORATION Road transform leading to a processing power equivalent to approximately 650000 equivalent correlators The minimum sensitivity in acquisition is 145 dBm obtained using coherent and non coherent integrations Thanks to the massive parallelization and the assistance TTFF at 145 dBm is still below 3 seconds Positioning Algorithms The OBU hosts positioning algorithms designed by NSL to provide high accuracy availability and integrity through exclusion of outlying measurements and provision of quality metrics horizontal protection levels or HPLs Numerous positioning algorithms and outlier detection strategies are being investigated These include consistency checks applied to raw measurements and computed positions and receiver autonomous integrity monitoring RAIM EGNOS corrections are applied to improve accuracy and integrity indicators user differential range error indices are used as coarse fault detection barrier Consistency checks on measurements include differencing pseudoranges between epochs and checking that this rate is below detect and exclude outliers before measurements enter the Checks at this stage identify if speeds are within expected ranges for the application and whether height changes are reasonable for example The RAIM algorithm is based on the maximum normed residual method For the detection procedure the test statistic is calculated based on weighted sum of the squares of the residuals This test statistic undergoes a globaltest chisquare distribution and is tested against thresholds that are computed based on the probability of false alarm Pfa and degrees of freedom number of measurements minus number of unknowns The exclusion procedure is based on an outlier detection technique also known as data snooping which is based on normed residuals and applied within the range domain This technique uses measures of internal and external reliability where the internal reliability gives estimates of how reliable the outlier detection procedure is ence of an outlier normed residual is tested against a critical value based on the normal inverse cumulative distribution which in turn depends on the Pfa and a decision is made on whether or not to exclude measurements Having performed fault detection and fault exclusion until no further outliers are found an HPL is calculated This is the maximum horizontal position error that is guaranteed by the algorithm not to be exceeded in accordance with the required probabilities of missed detection and false alarm HPL is a function of visible satellites expected error characteristics and user geometry Measurements which have been screened using the RAIM FIGURE 5 Journey record view from Nottingham test fault detection and exclusion are then processed in a Kalman Within the project many alternative algorithms and con snapshot mode to screen measurements entering the Kal that consider signal to noise ratios SNR as well as satellite elevations are also being used This combined weighting is useful in reducing the impact of measurements affected by multipath in urban areas where simple elevation dependent models are often not applicable The ultimate aim is to produce a robust GNSS positioning solution optimized for RUC in urban areas that balances the requirements of providing high availability with high integrity Test Methodology The SIGNATURE end to end solution was tested in a series 2010 Trials took place in a range of operating conditions from rural areas with open skies to dense urban environments In all trials assistance data was provided from the service center in Nottingham with messages tailored for the designated test area The OBU recorded real time position records can be sent back to the service center over a GPRS allowed alternative solutions to be applied to the original datasets in post processing The position solutions were assessed through comparisons with high accuracy GNSS reference solutions provided by additional onboard equipment and through processing with a map matcher NSLs Matchbox Each journey record from a trial was compared against the known reference journey record to determine charging availability accuracy and integrity GPS World March 2011 www gpsworld com 32