基于动态测量不确定度理论的最佳线性数据融合算法

Optimal linear multi-sensor data fusion algorithm based on dynamic uncertainty theory

本文提出一种基于动态测量不确定度理论的最佳线性数据融合算法。本算法构建过程中充分考虑了多传感器系统的动态特性,可用于处理工作环境复杂、性能可能随时间发生变化的多传感器系统的测量数据问题。同时,本算法具有实时监控各传感器性能和实时检出并剔除失效传感器的能力;测量不确定度的引入,使实时评价融合结果的质量成为可能;通过采用分级融合算法,又使得由截断误差引起的测量不确定度明显减小。

An optimal linear data fusion algorithm based on dynamic uncertainty theory is presented in this paper. It is suitable for the data processing of dynamic systems whose performances may change with time. It can monitor the sensor＇s performance and pick out the failure sensors in real-time. Because the adoption of dynamic measurement uncertainty, the algorithm can estimate the quality of the data fusion result in real-time. In addition, the hierarchical fusion algorithm can greatly reduce the measurement uncertainty caused by the truncation in digital computation.