非线性参数估计在超短基线安装校准中的应用

Nonlinear parameter estimation for calibration and installation of an ultra-short baseline system

比较了各种非线性参数估计循环解法在超短基线安装误差校准中的性能,并提出一种改进算法,使用循环算法进行估计,然后使用非线性模型参数估计的一种非循环(直接)解法进行修正.构建了安装校准问题所需的观测方程;分别实现牛顿法和信赖域法、最速下降法、高斯牛顿法和改进的高斯牛顿法、循环求解观测方程等,分析各解法的性能;并与非线性模型参数估计的非循环算法进行了比较,提出循环算法和非循环算法结合的改进算法,给出算法流程,并利用海试数据处理结果验证方法的有效性.

The performance of non-linear parameter estimation methods for the installation and calibration of ultra-short baseline（USBL） arrays was compared.An improved algorithm was proposed,where the parameters were estimated using a loop algorithm and then a non-cycle（direct） method was used to revise them.The equation of observation needed to solve the calibration problems for installation was found.Newton＇s method,the trust region method,steepest descent method,Gauss-Newton method and the improved Gauss-Newton method were implemented to solve the equation and their performance was analyzed and compared with non-cyclic algorithm for non-linear model parameter estimation,and then an improved algorithm that combines with a non-cyclic algorithm and a loop algorithm was given.The algorithm＇s procedures were discussed,and then sea trial data was used to verify its efficiency.