An Instantaneous Ambiguity Resolution Procedure Suitable for Medium-Scale GPS Reference Station Networks Horng-Yue Chen School of Geomatic Engineering, the University of New South Wales, Sydney, NSW 2052, AUSTRALIA BIOGRAPHY Horng-Yue Chen received a B.Sc. and M.Sc in Surveying Engineering (in 1984 and 1990 respectively) from the National Cheng Kung University, Tainan, Taiwan. He is currently a Ph.D. student in the School of Geomatic Engineering, The University of New South Wales, where his research is focussed on rapid GPS techniques for geodetic survey applications. ABSTRACT There is a trend for the establishment of regional-scale GPS permanent receiver networks, for a variety of applications including to support high accuracy, carrier phase-based positioning for surveying and precise navigation. When implemented in real-time, GPS users located within the region enclosed by multiple GPS reference stations can precisely position by using, for example, the 'correction terms' generated and transmitted by the reference station network. For such a configuration one of the major challenges is that the integer ambiguities have to be resolved during the real- time processing of the reference network data in order to ensure the generation of the carrier phase corrections, even when the reference receivers are many tens of kilometers apart. Due to the presence of distance dependent errors in the double-differenced data (principally the ionospheric and tropospheric delays) reliable instantaneous (single epoch) ambiguity resolution is difficult in the case of medium-scale reference networks (defined here as where the reference stations are typically in the range 50-100km apart). In practice the ambiguities between the reference stations could be correctly fixed during so sort of 'initialization' procedure. However, the main challenge is to keep on fixing ambiguities instantaneously (or with minimum delay) when tracking to a satellite experiences cycle-slips or a long data gap, or when a new satellite rises above the horizon. In this paper, a three-step strategy is proposed for overcoming these challenges which is suitable for implementation in real-time. In the first step the high correlation of the atmospheric delay between adjacent epochs is used to assist the cycle-slip recovery and ambiguity resolution. Then these atmosphere models are used in a predictive manner in the double- differenced observations, on an epoch-by-epoch and satellite-by-satellite basis. Finally, these atmosphere models are applied in a real-time kinematic data processing algorithm to fix the ambiguities during situations when there is a long data gap, or when a new satellite rises between the reference stations. Test data from GPS reference stations spaced 80km apart were used to evaluate the algorithm. The results indicate that the proposed algorithm can provide reliable integer ambiguities, very quickly, in the case of medium-scale reference networks. INTRODUCTION Regional-scale GPS station networks have been established to support multiple functions such as precision farming, surveying, crustal deformation monitoring, precise relative navigation, and so on. The GPS receivers involved in such applications fall into two main categories, those receivers that are part of the multiple reference station network, or the user (application) receivers. For real-time implementations, the multiple reference stations need to continuously make measurements, which must then be processed to generate some form of 'correction message' for transmission to users. Users could then use the correction message information to fix the integer ambiguities instantaneously. Hence, there are two major challenges: how to ensure that the multiple reference station network continues to generate the 'correction messages' without interruption, and how to use this correction message information in the user receivers to provide continuous precise positioning. In the case of medium-scale multiple reference station networks, resolving the ambiguities is difficult because of the presence of the significant residual atmospheric biases that remain when double-differences are taken between receivers several tens of kilometers apart. Therefore fixing the correct integer ambiguities in real- time (or near real-time) for the reference station data processing is a significant challenge. Li & Gao (1998), and Rabah & Leinen (1999), have suggested algorithms