一种基于卡尔曼滤波的定位解算性能评估新方法

对基于卡尔曼滤波的定位解算性能进行评估是提高卫星导航系统性能的有效途径。而传统的评估方法受人为主观性因素影响较大,提出了一种对数最小二乘模糊层次分析(FAHP)方法和模糊综合评价(FCE)方法相结合的性能评估方法。推导了模糊权重向量的唯一性确定条件,并将对数最小二乘FAHP方法确定的模糊权重向量去模糊化,使得权重向量可直接利用FCE方法进行加权综合,形成优势互补。以基于卡尔曼滤波的定位解算性能评估为实例,通过比较分析表明,所提出的对数最小二乘FAHP-FCE评估方法计算得到的权重向量优于基于程度分析的FAHP方法和基于先验规则挖掘的FAHP方法,使其更适合于对基于卡尔曼滤波的定位解算算法进行性能评估。

A Novel Evaluation Method of Positioning Resolution Performance Based on Kalman Filtering

The performance evaluation of positioning resolution based on Kalman Filtering is an effective way to improve the performance of satellite navigation system. As the traditional evaluation methods can be influenced subjectively, the paper proposes an efficacious evaluation method which combines fuzzy logarithmic least square AHP method with fuzzy comprehensive evaluation (FCE) method. The uniqueness determination conditions of fuzzy weights are deduced and the fuzzy weight vector determined by fuzzy logarithmic least square analytic hierarchy process (AHP) method is defuzzified. In this way, the unique and defuzzificated weights are available to synthesis for FCE method directly. Taking the performance evaluation of positioning resolution based on Kalman Filtering for a numerical example, based on some simulations carried out, we draw the conclusion that, the weight vector achieved by the proposed fuzzy logarithmic least square AHP-FCE method is superior to that by the fuzzy AHP method based on either extent analysis method (EAM) or apriori rule mining (ARM) when applied to performance evaluation of the positioning resolution based on Kalman Filtering.