混合整数最小二乘在GPS定位中的应用

GPS Positioning Based on Mixed Integer Least Square Estimation

目前针对载波相位双差观测模型中具有整型和浮点型两类参数的特点,常用的方法是先将混合整数最小二乘估计转换为整数最小二乘估计,并以此为准则搜索确定整周模糊度,然后回代解算基线向量。这里提出了一种基于混合整数最小二乘对两类参数同时估计的方法;推导了在未知参数具有先验信息的条件下混合整数最小二乘及整数最小二乘的参数估计准则。最后给出算例验证了基于混合整数最小二乘的参数估计的正确性和优越性。

There are two classes of parameters in the linear model for DD phase：integer parameters and real parameters, so the standard least square method can not be applied. Existing method for parameter estimation is always to reformulate the mixed integer least squares problem to the integer least squares problem, then integer parameters can be abtained by using integer least squares criterion, finally real parameters can be obstained by back substituting the original model. A new method based on mixed integer least square was proposed to solve integer and float parameter. Given prior information of the unknown parameters, the corresponding algorithms were derived. ,Some actual experiments were demonstrated to verify the mixed integer least square valid and advanced.