No abstract

In a recent electronic warfare (EW), electronic support (ES) systems suffer from the ambiguity problem that multiple radar types are reported instead of picking out the specific one. Hence, a radar scan pattern has been utilised as an important feature to improve the accuracy of threat identification. However, since false identifications cause immediate danger to friendly forces, the probability of false identification should be reduced as well as the increase in identification accuracy. To cope with this necessity, a new decision category entitled 'unidentified' is introduced based on the variance of the difference in peak-to-peak intervals. Firstly, the successive received signal strength is modelled for a given scan pattern, and then the effects of the position and movement of an ES receiver onto the identification accuracy are examined to analyse the tendency of false identifications. Simulations are included to confirm the effect of the proposed new feature parameter. By introducing the new parameter, on average 82% of the false identifications are classified into the new decision category instead of incorrectly being classified as a radar scan, whereas only 4% of the correct identifications as a raster scan are classified into the new category.

No abstract

No abstract

No abstract