具有不确定性平差算法

摘要

观测不确定性常常影响参数估计的有效性.将不确定度作为参数融入平差模型,可以有效地降低不确定性的影响.本文提出有界不确定性误差约束下,随机误差与不确定性误差平方和最小的平差准则,并给出了一个不确定性平差模型迭代算法.通过仿真实例,对不确定性最小二乘法与总体最小二乘法进行了比较.结果显示:在一定程度上,不确定性最小二乘方法的估计结果要略优于总体最小二乘方法,且在不确定性较大时,该方法有较好的适用性.%The uncertainty of observation often affects the validity of parameter estimation, and the effects of uncertainty can be reduced effectively by incorporating uncertainty into the adjustment model as an observation error parameter.An adjustment criterion is proposed under the bound constrain of uncertainty, in which the sum of squares of random error and uncertainty error should be minimized, and provided an iteration algorithm to solve the adjustment model.With simulation examples, the estimation results of uncertainty least-square method are compared with that of total least-square method.The results show that the estimation results of uncertainty least-square method are better than that of total least-square method to a certain extent and more applicable when uncertainty is greater.