论文部分内容阅读
数据关联是多传感器数据融合的关键技术之一,是对冗余数据进行融合处理的前提和基础。对关联算法的性能分析与评价是算法设计与选取的主要依据。传统的性能分析方法多数为基于仿真测试的事后评价方法,其分析结果与仿真场景密切相关,难以全面客观地反映算法性能。该文提出了一种从理论上分析关联算法性能的思路,首先寻找关联算法出错的边界条件,分析边界条件出现的概率,从而得到算法关联错误概率的期望。与传统性能评价方法不同的是,该方法揭示了关联算法的内在性能,以工程中常用的二维分配算法为例,分析了算法性能与系统误差的关系,为实际系统的算法设计提供了理论依据。仿真实验验证了理论分析的结论。
Data association is one of the key technologies in the multi-sensor data fusion. It is the premise and basis of the fusion processing of redundant data. The performance analysis and evaluation of the correlation algorithm is the main basis for algorithm design and selection. Most of the traditional performance analysis methods are ex-post evaluation methods based on simulation tests. The analysis results are closely related to the simulation scenarios and it is difficult to fully and objectively reflect the performance of the algorithm. This paper presents a way to analyze the performance of correlation algorithm in theory. First, we find the boundary condition of the association algorithm and analyze the probability of the boundary condition, so we can get the expectation of the probability of association error. Different from the traditional performance evaluation method, this method reveals the inherent performance of the correlation algorithm. Taking the two-dimensional distribution algorithm commonly used in engineering as an example, this paper analyzes the relationship between the performance of the algorithm and the system error, and provides the theory for the actual system algorithm design in accordance with. Simulation results verify the theoretical analysis.