Threat evaluation of enemy air fighters via neural network-based Markov chain modeling

Threat evaluation (TE) is a process used to assess the threat values (TVs) of air-breathing threats (ABTs), such as air fighters, that are approaching defended assets (DAs). This study proposes an automatic method for conducting TE using radar information when ABTs infiltrate into territory where DAs are located. The method consists of target asset (TA) prediction and TE. We divide a friendly territory into discrete cells based on the effective range of anti-aircraft missiles. The TA prediction identifies the TA of each ABT by predicting the ABT's movement through cells in the territory via a Markov chain, and the cell transition is modeled by neural networks. We calculate the TVs of the ABTs based on the TA prediction results. A simulation-based experiment revealed that the proposed method outperformed TE based on the closest point of approach or the radial speed vector methods.