Efficient Jamming Identification in Wireless Communication: Using Small Sample Data Driven Naive Bayes Classifier

The identification of malicious jamming pattern is significant for wireless communication system to adopt targeted anti-jamming approaches. Efficient identification of malicious jamming patterns requires sufficient sample data and computing resources for the complex electromagnetic environment. However, with the use of portable communication equipment and emergency communications, it is difficult to ensure adequate sample data and computing resources. Against this background, this letter focuses on the jamming identification based on a small sample data-driven Naive Bayes classifier. The main idea is to obtain approximate conditional feature probability via data augment and kernel density estimation (KDE) from a small sample labelled data and construct Naive Bayes classifier, so as to form a fast jamming recognition method. Simulation results show that compared with the methods using decision tree or simple neutral network, the proposed scheme achieves a better average accuracy between six jamming patterns, as well as low-complexity.