TY - GEN
T1 - A k-means and GMM-based fusion and detection algorithm against FDI attacks on remote estimator
AU - Hua, Jinxing
AU - Hao, Fei
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023.
PY - 2023
Y1 - 2023
N2 - This paper presents a K-means and gaussian mixturemodel (GMM) based detection and fusion algorithm for a multisensory cyber physical system (CPS). In the considered system, part of measurement channels may suffer from false data injection (FDI) attacks, which would deteriorate the estimation performance of the CPS. To handle this, a novel detection and fusion algorithm is proposed to eliminate compromised sensors and fuse safe sensors. Firstly, the K-means algorithm is utilized to get rid of severely biased sensors and the GMMalgorithm is subsequently adopted to further detect sensors screened by the K-means algorithm. Moreover, a more computationally efficient sequential Kalman filter is used at the remote estimator side, and the detection and fusion algorithm based on K-means andGMMalgorithms is derived in the framework of the sequential Kalman filter. In addition, the recursion of the estimation error covariance is recalculated in the presents of attacks. Finally, the effectiveness of the detection and fusion algorithm is verified by a simulation example of an unmanned ground vehicle (UVA).
AB - This paper presents a K-means and gaussian mixturemodel (GMM) based detection and fusion algorithm for a multisensory cyber physical system (CPS). In the considered system, part of measurement channels may suffer from false data injection (FDI) attacks, which would deteriorate the estimation performance of the CPS. To handle this, a novel detection and fusion algorithm is proposed to eliminate compromised sensors and fuse safe sensors. Firstly, the K-means algorithm is utilized to get rid of severely biased sensors and the GMMalgorithm is subsequently adopted to further detect sensors screened by the K-means algorithm. Moreover, a more computationally efficient sequential Kalman filter is used at the remote estimator side, and the detection and fusion algorithm based on K-means andGMMalgorithms is derived in the framework of the sequential Kalman filter. In addition, the recursion of the estimation error covariance is recalculated in the presents of attacks. Finally, the effectiveness of the detection and fusion algorithm is verified by a simulation example of an unmanned ground vehicle (UVA).
KW - Cyber physical systems
KW - False data injection attacks
KW - Gaussian mixture model
KW - K-means algorithm
KW - Sequential kalman filter
UR - https://www.scopus.com/pages/publications/85175058748
U2 - 10.1007/978-981-99-6882-4_12
DO - 10.1007/978-981-99-6882-4_12
M3 - 会议稿件
AN - SCOPUS:85175058748
SN - 9789819968817
T3 - Lecture Notes in Electrical Engineering
SP - 151
EP - 161
BT - Proceedings of 2023 Chinese Intelligent Systems Conference - Volume II
A2 - Jia, Yingmin
A2 - Zhang, Weicun
A2 - Fu, Yongling
A2 - Wang, Jiqiang
PB - Springer Science and Business Media Deutschland GmbH
T2 - 19th Chinese Intelligent Systems Conference, CISC 2023
Y2 - 14 October 2023 through 15 October 2023
ER -