Method for operating mode identification of multi‐function radars based on predictive state representations

Abstract

Operating mode identification of multi‐function radars (MFRs) is critical in radar threat evaluation and electronic jamming decisions. The hidden Markov model (HMM) in conventional methods relies on the MFR prior knowledge and cannot accurately describe the MFR signal sequence with regularity. An identifying method is presented that is based on predictive state representation (PSR) models without the help of prior information, which is unlikely to be obtained. Simulations of synthetic MFR signals are used to demonstrate the applicability of the novel method and to summarise the training parameter setting principles, which are of practical significance in applications of the method. The PSR‐based method is shown to be more effective than the HMM‐based methods, in particular when applied to the radar signal sequence with unknown regularity. The method also contains a novel approach to reducing the dimension of the system‐dynamics matrix, which enables the algorithm to perform well in conditions of heavy noise. Simulation results attest to the validity of the proposed method. With the proposed method, the operating modes of radars can be identified accurately, which is significant in supporting adaptive countermeasures against MFRs.