Recognition of Multifunction Radars Via Hierarchically Mining and Exploiting Pulse Group Patterns

Abstract

Recognition of multi-function radar (MFR) is an open problem in the field of electronic intelligence. Parameters of MFR pulses are generally agile and difficult to distinguish statistically. A prospective way to realize credible MFR recognition is mining and exploiting more distinguishable high-dimensional patterns buried in pulse groups, which may be designed for implementing infrequently-used radar modes such as target tracking. A high-dimensional pattern is defined according to the agile range and switching law of sequential pulse repetitive intervals (PRI) within a pulse group. This paper establishes deep recurrent neural networks (RNN) to discriminate and coarsely cluster different pulse groups hierarchically with respect to their sequential structures. Afterwards, RNN-based classifiers are trained to extract and exploit features within different pulse group clusters. Distinct degrees of confidence are then attached to these classifiers to indicate the discriminabilities of the corresponding pulse group clusters. The pulse group clustering and classifying models are finally cascaded to form an integrated classification model, which mines distinguishable patterns from sequentially arriving pulse groups of the same radar and accumulate them to realize MFR recognition. Simulation results demonstrate the much improved performance of the proposed method over existing counterparts in different scenarios.

Pulse Deinterleaving for Multifunction Radars With Hierarchical Deep Neural Networks

“…) do (1) Initialize the qth deinterleaving process: begin with the first pulse group x 1 of the interleaved pulse train, use encoder E to transform this pulse group to a vector e 1 according to (12), use classifier C to judge the corresponding radar type k according to (24), and select the corresponding pulse group prediction model R k . Denote the already deinterleaved pulse train as m q = x 1 , the number of sorted pulse groups as j = 1, the index of the first pulse in m after m q as n.…”

“…Intra-and inter-pulse parameters have been exploited to address various radar reconnaissance problems, such as radar recognition [9]- [12]. For the pulse deinterleaving problem concerned in this paper, time-of-arrival (TOA) is the most widely used parameter [13]- [16].…”

Multi-attribute overlapping radar working pattern recognition based on K-NN and SVM-BP

“…Paper on radar working recognition mainly divides radar working mode into four typical working: speed search, ranging while searching, scanning while tracking, and single target tracking [1]. For example, [26][27][28][29] (2,2) num r (2,3) num ⋯ r (2,N) num r (3,2) num r (3,3) num ⋯ r (3,N) num ⋮ ⋮ ⋱ ⋮ r (N,2) num r (N,3) num ⋯ r (N,N) num ⎤ ⎥ ⎥ ⎥ ⎦ side's disability to assess accurately the status of the enemy aircraft in air combat and even miss the aircraft. It has an irreversible impact on the air combat and the safety of the aircraft.…”

“…The recognition of radar working patterns is one of the main components of current cognitive EW. In the current battlefield, traditional radar has been gradually replaced by multi-functional radar (MFR) with multiple working modes and higher flexibility [2]. For example, airborne air-to-air radar functions in searching, tracking, taking images and making guidance systems.…”

Automatic pulse repetition pattern reconstruction of conventional radars

“…However, this method requires prior information about the high-dimensional PRI pattern of the radar pulse train, which is hardly available based on existing radar signal analysis methods. Deep learning-based radar signal deinterleaving [13] and recognition [14] methods also extract and utilize the pulse repetition pattern implicitly. However, the pattern is hidden in the parameters of deep neural networks, which is very unintuitive for human beings.…”