关键词: 辐射源个体识别, 脉内细微特征, 模糊函数, 零点切片, 特征优化

Abstract:

A novel specific emitter recognition algorithm is proposed to mitigate the computation problem and instability of kernel points ranking in the ambiguity function optimization method. It extracts the zero-slice of the ambiguity function as an intermediate feature set and employs the Direct Discriminant Ratio (DDR) criterion to rank the kernel points along the obtained slice. The resulting scheme not only improves the recognition accuracy greatly, but also preserves the most discriminant features of the individual emitter efficiently and avoids the “out of memory” problem in large whole-plane optimization. Experiments on both synthetic radar data from the U.S. Naval Research Laboratory and real radar emitter data demonstrate the feasibility and effectiveness of the proposed method.

Key words: specific emitter recognition, intra-pulse fine features, ambiguity function, zero-slice, feature optimization