Abstract: The slice of Ambiguity Function Main Ridge（AFMR） can describe the structural information of the radar signals more completely, and its corresponding characteristic parameters can be as the effective supplement of five classical signal sorting parameters. But the search of AFMR slice needs large amount of calculation, which is against the real-time radar signal deinterleaving. A fast extraction method of AFMR slice which based on the improved particle swarm optimization has been proposed in this paper. This method effectively avoids the algorithm into the local optimal and improves the convergence speed by uniform initializing population, stochastic inertia weight and natural selection technology. The experimental results show that the proposed method can work reliably under the SNR not less than 6 dB, also it can obtain more accurate AFMR slice and increase the search speed significantly at the same time. The results confirm the feasibility and effectiveness of the proposed method.

Key words: radar emitter, signal deinterleaving, ambiguity function main ridge, particle swarm optimization

摘要： 模糊函数主脊切面能较为完整地描述雷达辐射源信号的结构信息，其特征参数可作为信号分选经典五参数的有效补充。但搜索信号模糊函数主脊切面的计算量较大，不利于雷达信号的实时分选。提出一种模糊函数主脊切面特征快速提取的改进粒子群优化方法，该方法通过均匀初始化策略、随机惯性权重与自然选择等技术，有效避免了算法陷入局部最优并显著提高收敛速度。实验结果表明，所提方法能可靠地工作在SNR不低于6 dB的情况下，在显著提高搜索速度的同时还可获得更为精确的模糊函数主脊切面，进一步证实了方法的可行性和有效性。

关键词: 雷达辐射源, 信号分选, 模糊函数主脊切面, 粒子群优化