A Segmented Template Optimization Using the Frechet Distance
Peibo Chen, Ke Xu, Gang Li, Jianwei Wan · 2016
As the modern radar system become increasingly multi-functional and sophisticated, the specific emitter identification (SEI), recognizing the different radar devices of the same type, seems to be one of the most important tasks in electronic warfare. The main operation of the SEI is utilizing a distance function to measure the distinction between the testing feature vector and each template vector. In this paper, the time-domain waveforms (pulse envelope and instantaneous frequency) and frechet distance, which have outstanding performances in pattern classification, are specified as the fingerprint features and distance function respectively. Unfortunately, the high-dimensional fingerprint features and computational complexity of the frechet distance have brought a tremendous workload on the template optimization. In order to overcome this difficulty, an improved method, the segmented template optimization algorithm, is proposed in the paper.