A novel segmentation approach for work mode boundary detection in MFR pulse sequence

Understanding and analysis of multi-function radar (MFR) work modes play a key role in radar countermeasures. Many methods have been proposed for recognizing the work modes. However, segmentation of pulse sequence is prior to the mode recognition as the mode transition boundaries are not known in advance for the whole sequence. To solve the boundary detection problem, an unsupervised segmentation method based on recurrence plot (RP) and singular value decomposition (SVD) is proposed in this paper. The method utilizes RP to reveal the characteristic change of the pulse sequence, and makes a quantitative analysis through SVD to extract the main information of work mode. Then, discrepancy measure is employed to detect the boundary, thus determining whether or not the mode transition has occurred. The proposed method can recognize the transition boundaries at different granularity levels. Experimental results show that our method achieves better performance than existing segmentation methods and is robust to non-ideal conditions.