부산대학교 도서관

When detecting radar targets, we need to infer the specific type of radar the target is by obtaining the feature data, such as a certain type of aircraft loaded radar. In the past, this work was usually recognized manually, depending on the staff’s familiarity with the feature data of radar targets. Moreover, the values of the feature data for different types of targets are sometimes very similar. All these circumstances described above have resulted in the low accuracy and efficiency of recognition. Therefore, this paper develops a radar emitter identification system, which uses one of the most well-known and widely being used machine learning algorithm K-nearest neighbor (KNN) to assist the staff to recognize the target. According to the feedback of users, the system has achieved excellent results.