Non-cooperative target recognition in multistatic radar systems

The ability to recognise quickly and reliably non-cooperative targets is a primary issue in modern radar systems. The use of multistatic radars has already been demonstrated to improve sensibly the recognition capacity of a radar system, through the exploitation of its inherent spatial diversity, given by the presence of multiple observation channels. This calls for the development of suitable fusion strategies for efficiently combining the signals arising from multiple channels of the network. In this study, the authors define a multistatic classification algorithm where the information on the target class is provided by the sensors of the system and the final decision is made using a fusion rule that combines the decisions coming from each channel of the radar network. A suitable selection is made to favour the channels with better performance. The numerical analysis demonstrates the superior performance of the proposed algorithm over the classical fusion rule, that takes into account only the effect of the energy path loss, in terms of probability of correct classification.