Recognition of MFR Based Motif Discovery

Multi-function radar (MFR) has pulse-level parameter agility, and under the scheduling of complex procedures, it can adjust resource allocation in an optimal way to perform multiple tasks in different directions. The agility of the parameters makes it difficult to describe MFR statistically, and traditional statistical description methods based on signal characteristics are difficult to support the needs of MFR recognition tasks. Another potential MFR description method is using the high-dimensional temporal pattern hidden in pulse stream, which is represented by the pulse repetition interval (PRI) switch rule and value range. This paper uses motif mining algorithm to find high-dimensional repeated patterns in MFR pulse stream, and the key PRI sequence of MFR is obtained through further screening. Then, multiple key sample sequence matchers are integrated into a matching degree generator, which is used to output the matching degree between the MFR to be identified and the reference MFR. Simulation results show that compared with the deep learning method, it has better performance and stronger interpretability.