The University of British Columbia

We consider the electronic warfare problem of decoding the internal state of a multi-function radar at a given time, when provided with incomplete observations of the radar emissions and some prior knowledge of how the radar functions. The internal state of multi-function radar can provide an indication of where it stands in its target acquisition process and in how much time a weapon can be triggered. Therefore, estimation of the internal state of the radar is useful to assess how much of a threat it constitutes. In this paper a mathematical framework to model how radars function is proposed. This framework is based on the concept of generalized semi-Markov processes. In the case of radars structuring their emissions around a clock, this model reduces to a Markov-modulated Markov process. In the latter case, the problem of decoding the internal state of the radar given incomplete observations can be solved by using a discrete-time hidden Markov filter. The practicality of the proposed approach is discussed and illustrated by means of a hypothetical multifunction radar example.