一种组合导航信息融合算法品质评估方法

针对组合导航信息融合算种类繁多,给决策带来了优选的难题,提出了一种信息融合算法品质评估方法.首先对算法品质进行了全面的分析,提出了一套反映算法的精度、实时性、稳定性及可靠性的评估指标体系,并给出了定义及数学模型；然后研究了多指标综合评判问题,针对传统模糊评判法确定的指标权重客观性差,确定的评判等级分辨率不强等问题,采用熵技术修正权重,并将评判等级进行区间量化,基于此提出了一种改进的模糊综合评判模型.评估实验结果表明,所提出的指标体系能够比较全面、合理的评价算法的品质,改进的评判模型进一步提高了评估结果的客观性和分辨率,所提出的评估方法可以有效用于信息融合算法的品质评估.

Quality evaluating method of integrated navigation information fusion algorithm

To solve the optimum seeking problem for there are various algorithms of integrated navigation information fusion,a quality evaluating method of information fusion algorithm is presented.Firstly,the paper comprehensively analyses the algorithm quality,and proposes an evaluating index system relating to precision,real-time quality,stability and reliability of the algorithm,and simultaneously develops the corresponding definitions and mathematical models.Then the multi-index synthetic evaluation method is analyzed.The traditional fuzzy evaluation method is not so objective when confirming the weight of index and has a low resolution ratio when confirming evaluation matrix.To overcome these shortcomings,an improved fuzzy synthetic evaluation model is proposed,in which the index weight is amended objectively by using entropy technology,and the evaluation grade is quantified using interval value.The evaluation experiment results show that the presented index system can evaluate the quality of algorithm comprehensively and reasonably,and the improved evaluation model enhances the objectivity and resolution ratio of evaluation results,and the proposed evaluation method is valid in evaluating the quality of information fusion algorithm.