一种基于改进PSO的随机最大似然算法

发布时间：2018-03-25 21:44

本文选题：波达方位估计　切入点：粒子群优化算法　出处：《电子学报》2017年08期

【摘要】：随机最大似然算法(Stochastic Maximum Likelihood,SML)具有优越的波达方位(Direction-of-Arrival,DOA)估计性能,但SML解析过程较高的计算复杂度限制了该算法在实际系统中的应用.针对SML计算复杂度高的问题,提出一种低复杂度的粒子群优化算法(Particle Swarm Optimization,PSO),解决了传统PSO算法中粒子数多和迭代次数多的双重缺点.首先,根据天线获得的信号,将旋转不变子空间法(Estimation of Signal Parameters via Rotational Invariance Techniques,ESPRIT)求得的闭式解作为DOA的预估计值,同时计算系统此时的信噪比以及SML在此信噪比下的克拉-美罗界(Cramer-Rao bound,CRB).然后,根据DOA预估计值和当前CRB值在SML最优解的近邻范围内确定较小的初始化空间,并在该空间初始化少量粒子.最后通过设计合适的惯性因子w,使粒子以合理的速度搜索最优解.实验结果表明,改进PSO算法所需的粒子个数和迭代次数大约是传统PSO算法的1/5,降低了SML的解析复杂度,计算时间是传统PSO算法的1/10,因此在收敛速度上也有显著的优势.
[Abstract]:Stochastic Maximum Likelihood-SMLs have excellent DOA estimation performance, but the high computational complexity of the SML analysis process limits its application in practical systems. A low complexity particle swarm optimization algorithm (PSO) is proposed, which solves the double disadvantages of large number of particles and many iterations in the traditional PSO algorithm. Firstly, according to the signal obtained from the antenna, the particle swarm optimization algorithm (PSO) is proposed. The closed solution obtained by the Signal Parameters via Rotational Invariance technique Esprit is taken as the pre-estimate value of DOA, and the SNR of the system and the Cramer-Rao boundary of SML at this SNR are calculated simultaneously. Then, According to the DOA pre-estimate value and the current CRB value, the smaller initialization space is determined in the nearest neighbor range of the SML optimal solution. A small number of particles are initialized in this space. Finally, by designing an appropriate inertia factor w, the particle can search the optimal solution at a reasonable speed. The experimental results show that, The number of particles and the number of iterations needed to improve the PSO algorithm are about 1 / 5 of that of the traditional PSO algorithm, which reduces the analytical complexity of the SML algorithm, and the computational time is 1 / 10 of the traditional PSO algorithm, so it also has a significant advantage in the convergence speed.
【作者单位】：中国石油大学(华东)信息与控制工程学院;中国石油大学(华东)计算机与通信工程学院;
【基金】：国家自然科学基金(No.61305012)
【分类号】：TN911.7

【相似文献】