A Soft-Decision Weighted CNN for Specific Emitter Identification in Aeronautical Monitoring System
Zhiheng Guo, Yuqiao Huang, Yuxin Li, Zhihao Chen, Xiang Chen · 2022 3rd Asia-Pacific Conference on Image Processing, Electronics and Computers · 2022
Specific Emitter Identification (SEI) is a hot topic in the field of aeronautical communication such as aviation safety. In this paper, we proposed a soft-decision weighted CNN model to make SEI in the field of special aeronautical monitoring system, i.g. Automatic Dependent Surveillance-Broadcast (ADS-B). Specifically, with the combination of statistical analysis, machine learning and neural network, the proposed model consists of statistical judgment, message clustering and network classification via CNN. The problem is that there are one hundred aircraft signals to be identify with limited data. Thus, we first calculate the soft decision probability and establish the radio frequency fingerprint (RFF) library via statistical analysis. Then, the K-Means is used to cluster the difference types of message and utilize the space characteristics of the aircraft such as speed, position and so on. At last, CNN is introduced to the model with the input of each RFF to make the classify for each cluster. We make the final identification by combining the soft decision probability and CNN classification result. The numerical result demonstrates that the proposed model can achieve performance with ADS-B data and the precision of identification is more than 90%.