Evaluation of a model based data fusion algorithm with multi-mode OTHR data

Abstract

Australia's over-the-horizon-radar system, the Jindalee Operational Radar Network, is due to be commissioned in 1997. When operating it is possible that associated tracks of one aircraft may arise from multiple propagation paths and multiple radars. Errors in estimating the propagation path result in errors in estimating the position of the aircraft. Also, the structure of the ionosphere typically supports several propagation modes and it may not be certain which mode caused each track. Generally, association of data from multiple sensors requires that the data be aligned in all its coordinates. If the propagation modes are uncertain, then alignments are compromised. This paper presents the results of applying a previously presented multi-mode multi-radar fusion algorithm to the experimental Jindalee Facility Alice Springs (JFAS) data. The algorithm, which is briefly summarised, uses a simple model of propagation to convert between radar coordinates and latitude-longitude coordinates. The algorithm has been applied to ten days of JFAS data. The results are shown to be comparable to operator performance, with indications that more improvements can be made. The difficulties of evaluating data fusion algorithms are discussed and it is concluded that development of data fusion algorithms is best done with both simulated and synoptic data.<>