Ensembles of Predictive Radar Models for Electronic Intelligence
Sabine Apfeld, Alexander Charlish · 2022 IEEE Radar Conference (RadarConf22) · 2022
Predictive radar models allow for the anticipation of the next emissions, which facilitates many tasks in an electronic intelligence context. If various radars have to be considered during a mission, several models need to be combined into a multi-model system or an ensemble. This paper examines three ensemble architectures in different configurations, which either contain radar models composed of Markov chains or Long Short-Term Memory networks. All approaches are based on a hierarchical emission model that understands radars as systems that speak a language. The ensembles' prediction accuracies are evaluated with a simulated airborne multifunction radar using three resource management methods of varying complexity. The evaluations consider several conditions, which include ideal and corrupted data, as well as different signal lengths.