超声回波参数的蚁群算法估计  

Parameters estimate of ultrasonic echo based on ant colony algorithm

作　　者：周方[1] 张小凤[1] 张光斌[1] 

机构地区：[1]陕西师范大学物理学与信息技术学院,陕西西安710062

出　　处：《陕西师范大学学报（自然科学版）》2012年第2期35-40,共6页Journal of Shaanxi Normal University：Natural Science Edition

基　　金：陕西省自然科学基金资金项目(2009JM1004;2010JM1008);中央高校基本科研业务费专项资金项目(GK200902056;GK201001008);西安市科技计划资助项目(CXY1009(2))

摘　　要：将用于连续函数优化的蚁群算法应用于超声回波参数估计.根据高斯回波模型,给出了用于超声回波估计的蚁群算法的基本原理和参数估计步骤.通过数值仿真,对不同信噪比条件下超声回波参数中的到达时间和中心频率进行了估计.结果表明:该方法不依赖于初始值的选取,可在较大范围内搜索,得到全局最优解,避免了高斯-牛顿迭代法模拟超声回波的缺陷,即解决了选择的初始值只有在真实值附近才能迭代出最优解的问题.通过与高斯-牛顿迭代法和模拟退火算法的比较,证明了蚁群算法是一种有效的回波参数估计方法,具有较高的精度.Presents a method of the parameter estimation of ultrasonic echo based on ant colony algorithm which used for continuous function optimization.According to Gauss echo model,the basic principle and parameter estimation procedures of ant colony algorithm are given.The arrival time and the center frequency are estimated on the conditions with different SNRs（signal to noise ratio） through the numerical simulation.The simulation results show that this method does not depend on the selection of the initial values and can search in a wider scope to obtain the optimal solutions,hence the flaw of the simulation of ultrasonic echo with Gauss-Newton method,which is the Gauss-Newton iteration only can iterative the optimal solutions when the selected initial values are in the vicinity of the true values is avoided.Through the analysis and comparison with Gauss-Newton iteration and simulated annealing,the ant colony algorithm is an effective method to estimate the parameters of ultrasonic echo,and has a higher precision

关 键 词：超声回波模型 蚁群算法 高斯-牛顿法 超声检测 

分 类 号：O657.5[理学—分析化学]