多传感器数据融合评估系统的设计与实现

Design and Implementation of Evaluation System for Multi-Sensor Data Fusion

多传感器数据融合技术现已成为研究热点,但对它的研究往往缺乏必要的仿真手段和评估手段。为不断验证和改进融合系统的融合算法,设计并实现了一个多传感器数据融合评估系统:首先建立了评估系统的模型;接着对评估指标,按精度、相关性和处理能力三方面内容,给出了详细定义和计算方法;并针对评估实现中的关键问题,给出了基于最近邻域的相关技术、基于自适应均值门限判断的正确航迹判断技术和基于Matlab的可移植编程技术等解决方法。仿真实验表明,评估指标设计合理,评估系统切实可用。

Based on a proposed model of multi-sensor data fusion evaluation system, this paper gives the detailed definitions and computation approaches for three kinds of evaluation indexes, i. e. , precision index, correlation index and capacity index. Furthermore, three important techniques, e. g. , the nearest neighbor based correlation, correct track judging based on adaptive mean thresholds and scalable programming based on Matlab are provided to solve the key problems in the system realization. Finally, the numerical experiments are also provided to show that the proposed evaluation indexes are effective and the established system is practicable.