Modern Methods of Radar Data Extraction and Track Management Systems

摘要

First we give an overview of the content of a dighalized target plot and the conflicting requirements of a track management system. This should find a reasonable compromise since a bijective plot-track-assignment does not exist. After this overview we discuss the new methods of radar data extraction that we have examined. We have studied the effectiveness of image processing algorithms and track before detect methods for plot formation. The latter have only been used up to now for offline evaluation because of the increased calculation time. In the development of multi-hypothesis tracking (MHT) we had several different approaches: We had to distinguish between the track initialization and firm track phases, have separated the two phases and developed an MHT for track initialization. We have also implemented an MHT for the firm track phase without cluster hypothesis in SDF (Sensor Data Fusion) in the F124 project. We have considered implementing an MHT with cluster hypothesis in our ADA software and have built demonstrators in ADA95 and in Java. Additional we have applied for a patent for a fully automatic sensor misalignment correction method which is able to correct misalignments of sensors in azimuth and range, and slowly changing systematic errors (bias). In the last section, conclusions, we give an overview of our future activities.