用于跳频分量选取的修正适应度距离比粒子群算法

跳频信号参数估计是跳频信号截获、干扰的前提,而传统Cohen类时频分析方法存在核函数选择的瓶颈.结合匹配追踪和智能计算的思想,将多峰函数粒子群优化算法引入跳频信号时频分析领域.在分析粒子适应度和粒子间距2个影响粒子搜索行为的关键因素的基础上,提出了基于改进的适应度-距离比测度的多峰函数粒子群优化算法,并应用于跳频分量自适应选取.该方法不需要跳频信号的任何先验知识和粒子群小生境参数的人为设置.理论分析和仿真结果表明,与基于环形拓扑结构、单一共享适应度信息的粒子群优化算法相比,算法成功率和参数估计精度进一步改善,该方法的邻域搜索机制和跳频分量选取具有可行性和有效性.

Modified fitness-distance ratio based particle swarm optimizer for selection of frequency hopping components

Parameter estimation of frequency hopping (FH) is the premise of interception and interference, while the traditional Cohen time frequency analysis (TFA) method has a bottleneck of kernel function selection. In this paper, combined with the matching pursuit and intelligent computation, particle swarm optimization (PSO) algorithm for multimodal function is introduced into the field of FH signal TFA. A PSO for multimodal function based on modified fitness-distance ration is proposed and used for FH component adaptive selection through analysis of two key factors of fitness and spacing which has impact on particle search behavior. This method does not need manual setting of particle swarm niching parameters and any prior knowledge about FH signals. Theoretical analysis and simulation results showed that it has the feasibility and effectiveness of neighborhood search mechanism and FH component selection, which the success rate and estimation precision were further improved, compared with the PSO with the ring topology and single fitness sharing.