An intelligent Small Area DGPS

Several techniques for the implementation of Differential Global Positioning System (DGPS) have been developed and demonstrated to significantly improve the positioning accuracy of the GPS system. The concept of Wide Area DGPS (WADGPS) has provided a practical approach for the development of a world wide network of the DGPS schemes. Current plans are to implement WADGPS under the auspices of Wide-Area Augmentation System (WAAS) in the United States of America and as the Future Air Navigation System (FANS) throughout the rest of the world under the sponsorship of the International Civil Aviation Organization (ICAO). The operations of the WAAS and FANS systems entail the set up of additional infra structure, and a commitment of significant financial and other resources. For small countries like Singapore, a preferred approach embodied in the concept of Small Area DGPS (SADGPS) has been suggested. The SADGPS aims to develop an autonomous DGPS controller which will eliminate the need for transmission of DGPS corrections in real time and thus provide a more robust as well as economical method to improve the performance of the basic GPS system. The investigation of SADGPS has been ongoing at the Nanyang Technological University in Singapore. Earlier, it was demonstrated that limited success for the SADGPS could be obtained by the application of constant corrections to the coordinates of points located within a small area. However, performance was sensitive to the brands of the commercial GPS receivers used. For some types of GPS receivers, significant improvement in accuracy was possible, whereas, for others no discernible improvement could be demonstrated. This paper presents the design of an intelligent SADGPS system based on the technology of artificial neural networks (ANNs). This SADGPS controller reduces the errors of basic GPS system below 10 meters and performs equally well for different brands of the GPS user equipment. To the author's knowledge this is the first published application of ANNs to DGPS.