Target Detection, Tracking and Threat Evaluation in Multi Sensor System using Machine Learning

Air defense is a crucial part of the defense system of any country. Most of the systems in air defense are equipped with multiple sensors for data acquisition. Once the target is detected it is very important to classify the target in order to carry out weapon allocation. This paper proposes a deep learning based SSD (Single Shot Detection) algorithm for the purpose of target detection and classification. This data is combined with data from radar sensor using Hungarian algorithm and then used for tracking. This paper uses IMM filter for the purpose of tracking. Detection, classification and tracking information is then passed to a Bayesian network for threat evaluation.