Real time detection and compensation of registration errors in a network of radars

This paper describes an algorithm designed to automatically compensate for the angular and positional misalignment of the members of a network of radars using nothing more than the operational data. The method is based upon the simple observation that within any given observation period a significant number of the target aircraft will take flight paths along lines of constant bearing. By using such sets of straight line motion, observed commonly by a pair of radars, it is possible to arrive at an estimate of the relative azimuthal bias between the two. By conducting this estimation for each radar in the network with reference to one nominated member, all the relative azimuthal offsets can then be computed. Once the rotational biases have been resolved it is then feasible to calculate any x,y offset in the actual position of the radars relative to their published coordinates using the same data. This paper is illustrated with a set of data gathered from five typical air defence radars during a thirty minute period of normal operation. A colour keyed latitude/longitude plot of all the radar reports is shown; the misregistration of the radars can be seen quite clearly. The reports from the radar designated as the reference are those shown in black. This is a good test data set in that, compared to the reference, one radar has only a slight offset, two have very substantial azimuthal bias and one of these is also located some tens of seconds adrift from its published location; finally one has only a limited area of coverage common to the designated reference.