一种基于记忆渐消因子指数加权的动态分布式传感器融合算法

A Dynamic Distributed Sensor Fusion Algorithm Based on Weighted Fading Factors Memory Index

由于新疆生物氧化提金预处理过程的工业现场易受到外界因素如昼夜温差、强风等干扰,传统测量方法往往存在较大的误差。为解决这个问题,①根据预处理过程,建立了传热机理模型。②针对该过程设计一个小范围传感器网络,提出了一种基于多连通融合结构的传感器分层融合结构。③数据处理过程中,引入一种基于渐消记忆指数加权的多重衰落因子调整预测误差协方差,提高基于扩展卡尔曼滤波(EKF)在一步预测中的有效性。④以各传感器的状态估计精度用作加权融合准则,通过添加动态加权因子来预测每个传感器的预测置信度。仿真实验的性能指标表明,该方法比传统的单传感器方法具有更高的全局精度,并能有效降低干扰噪声的影响。

Due to the interference of external factors such as temperature difference,strong wind and so on,the traditional measurement methods in the pretreatment process of biological oxidation gold extraction in Xinjiang often has large errors. In order to solve this problem,(1)according to pretreatment process,the heat transfer mechanism model is established.(2)A small range sensor network is designed for the process. A sensor hierarchical fusion structure based on multi-connected fusion structure is proposed.(3)In the process of data processing,a multiple fading factors based on weighted fading memory index to adjust the prediction error covariance to improve the effectiveness of EKF in one step prediction.(4)the state estimation accuracy of each sensor will be used as a weighting principle for the predictive confidence of each sensor by adding a weighting factor. The performance index of simulation experiments shows that the proposed method has higher global accuracy than the traditional single sensor method and can effectively reduce the influence of interference noise.