Naive Bayesian Classifier-based Collaborative Tracking and Recognition for Multi-radar Systems

This paper presents a naive Bayesian recognition and classification method based on extended Kalman filter, designed for multi-target tracking scenarios using multiple radars. The approach employs covariance intersection (CI) fusion to integrate data from diverse sources and obtains the target state, accordingly. A naive Bayesian classifier is trained with various features extracted from the target state, enhancing the classification process. The numerical findings reveal the algorithm’s robust performance in scenarios with multi-dimensional feature inputs, demonstrating its efficacy in classifying diverse target types.