Standalone Real-Time Positioning Algorithm Based on Dynamic Ambiguities (DARTS) Andrew Simsky, Septentrio BIOGRAPHY Dr Andrew Simsky holds a Ph.D. in physics from the University of Moscow (Russia). He is working as a senior GNSS scientist at Septentrio in Leuven, Belgium. His research interests include differential and standalone navigation algorithms and performance analysis of GNSS receivers. He previously worked on carrier-phase DGPS algorithms for airborne gravimetry at Sander Geophysics Ltd in Ottawa, Canada. ABSTRACT This paper presents an original navigation algorithm intended to push the accuracy of standalone positioning beyond the limitations imposed by GPS system biases (errors of broadcast orbits and clocks). The proposed algorithm is based on the processing of iono-free carrier- phase combination with floating ambiguities by the Kalman filter. Unlike traditional phase processing, ambiguities are essentially modeled as dynamic values intended to absorb long-term range errors, of which GPS system biases are the most essential. The model, which describes the behavior of ambiguities, has been optimized using both experimental and simulated data sets. The acronym DARTS reads as Dynamic Ambiguities Real- Time Standalone. The accuracy of DARTS is close to typical accuracies of code-based DGPS. Standard deviations for position are within the range of 1.0-1.3m for heights, and 0.6-0.8 m for horizontal coordinates. This level of accuracy has been proven for a variety of applications. DARTS has been designed having in mind its application in GNSS receivers. Its implementation for the new firmware version of Septentrios GNSS receiver, PolaRx2, is currently underway. INTRODUCTION After SA has been turned off in May 2000, single-point positioning algorithms using iono-free carrier phase processing with floating ambiguities have gotten significant attention. In [1] it has been shown that decimeter-level accuracy can be achieved when post- mission precise orbits and 30-sec clock corrections are used. However, such algorithms can be used only for post-processing, when precise IGS products become available. For real-time single-point applications, navigation algorithms are subjected to GPS system biases, i.e., errors of broadcast orbits and clocks, also called user range errors (URE). System biases are traditionally considered as an inevitable contribution to the error budget of single- point positioning and result in positional errors of 2-5 meters. The only way to improve the positional accuracy of non-assisted real time single-point positioning is to estimate GPS system biases in real-time and apply these corrections to GPS observations. This is exactly what DARTS is designed for. In order to estimate system biases and compensate for them in real-time, DARTS employs the concept of dynamic ambiguities. According to the traditional formulation of carrier phase processing, floating ambiguities are modeled as constant values and hence are expected to show converging behavior. In DARTS, ambiguities are modeled as non- constant, slowly changing values, and their expected behavior is to follow the long-term variations of range errors. The way of estimating system biases is, therefore, implicit: the variations of biases are absorbed by variations of dynamic ambiguities. The effectiveness of DARTS can be best appreciated by comparing it to DGPS. Indeed, in differential processing system biases are fully canceled (except for a small residual effect of orbit errors). With DARTS, only partial compensation of system biases can be achieved. It is shown further in this article that the accuracy of DARTS is comparable to the accuracy of DGPS, and therefore it can be concluded that a substantial portion of system biases is compensated by DARTS. In the following sections the principle of DARTS is described in detail. Optimization of DARTS is illustrated 1211 ION GPS/GNSS 2003, 9-12 September 2003, Portland, OR