Universal Kriging for Loran ASF Map Generation

Abstract

After a series of intentional Global Positioning System (GPS) jamming attacks impacted a large area of South Korea, the Ministry of Oceans and Fisheries of South Korea considers long-range navigation (Loran) and enhanced Loran (eLoran) as a maritime backup navigation system. Despite its robustness to signal jamming, the positioning accuracy of Loran/eLoran is lower than that of GPS. Because the signal delay due to the land path, which is called the additional secondary factor (ASF), is the largest unknown component of Loran/eLoran, it is necessary to account for temporal and spatial ASF errors to ensure high accuracy. The generation of ASF maps based on ASF survey data in a service area is the most convenient way to mitigate spatial ASF error, but the quality of ASF maps depends on the applied interpolation algorithm. It is desirable to generate high-quality ASF maps based on ASF measurements at only a few survey points, because extensive ASF surveys are expensive and time consuming and require considerable effort. This paper proposes kriging methods for satisfying this objective and shows their superior performance during a field test in Incheon, Korea. In particular, universal kriging with the proposed drift model showed a better performance than linear interpolation, inverse distance weighing, and ordinary kriging when the test vehicle was close to a coastline. The positioning accuracy with the ASF maps generated by the proposed universal kriging along a 5-km route during the field test was 25.24 m (95%). The land vehicle used for the test experienced significant signal-to-noise ratio (SNR) degradation owing to the noise caused by its engine. A vessel without such SNR degradation is expected to achieve higher accuracy.