China
Africa
Explore chapters and articles related to this topic
Cooperative Localization for Autonomous Vehicles Sharing GNSS Measurements
Published in Chao Gao, Guorong Zhao, Hassen Fourati, Cooperative Localization and Navigation, 2019
Khaoula Lassoued, Philippe Bonnifait
However, this work does not address the consistency/integrity of the estimates. The covariance intersection method represents a solution of the data incest problem since it manages the dependencies between estimates, and at the same time offers reliable confidence domains. The above-mentioned methods are all based on Bayesian approaches (EKF) or particle filter. Other fusion approaches can avoid the problem of overconvergence while guaranteeing results with reliable confidence domains, such as set-membership approaches [17–20]. This will be the following approach in this study, in particular the set-membership method based on interval analysis.
Multiple Sensor Estimation
Published in Bin Jia, Ming Xin, Grid-based Nonlinear Estimation and Its Applications, 2019
Covariance intersection is an algorithm to combine two or more estimates when the correlation between them is unknown. It provides a method to update the local estimate using the estimates of other sensors. The covariance intersection gives a common upper bound of the actual estimation error variances. It is robustness to unknown correlations. In addition, the fused estimate is consistent and the accuracy outperforms the single estimate (Li et al. 2015).
Essence of Distributed Target Tracking
Published in David L. Hall, Chee-Yee Chong, James Llinas, Martin Liggins, Distributed Data Fusion for Network-Centric Operations, 2013
Shozo Mori, Kuo-Chu Chang, Chee-Yee Chong
The covariance intersection (CI) method was introduced as a method of fusing two estimate-covariance pairs, (x^1, V1) and (x^2, V2) when the cross-covariance V12 of the estimation errors is not known or available. The CI approach is a heuristic approach to adjust the commonly used the simple weighting, i.e., the Speyer fusion rule (6.9), as () {VF−1x^F=αV1−1x^1+(1−α)V2−1x^2VF−1=αV1−1+(1−α)V2−1
Fusion estimation in clustering sensor networks under stochastic deception attacks
Published in International Journal of Systems Science, 2018
Haiyu Song, Zhen Hong, Hongbo Song, Wen-An Zhang
To deal with the fusion estimation problems with multiple distributed sensors, a series of fusion algorithms have been proposed, see for example, Refs. (Bar-Shalom, Li, & Kirubarajan, 2001; Dou, Ran, & Gao, 2016; Li, Zhu, Wang, & Han, 2003; Lin & Sun, 2017; Shen, Song, Zhu, & Luo, 2009; Song, Xu, & Zhu, 2014; Sun & Deng, 2004; Sun et al., 2017; Tan, Shen, Liu, Alsaedi, & Ahmad, 2017) and references therein. Note that in these references, the estimation error cross-covariances among local estimators need to be known exactly. However, in practical, the computation of cross-covariance matrices among local estimators may be very complex, or even the cross-covariance matrices cannot be obtained. To overcome these difficulties, the covariance intersection (CI) fusion method was proposed to fuse the sensor data with unknown correlations Julier and Uhlman (1997, 2009). The advantages of the CI fusion estimation are attractive. For one thing, the computation of the cross-variances can be avoided. For another, by using the CI fusion method, the FC has a better estimation accuracy than that of each local estimator. Recently, the CI fusion estimation algorithms have been further developed (Chen, Hu, Ho, & Yu, 2016; Deng, Zhang, Qi, Gao, & Liu, 2013; Deng, Zhang, Qi, Liu, & Gao, 2012) and also have been introduced into clustering sensor networks (Song et al., 2014; Zhang et al., 2015; Zhang et at., 2016).