TY - GEN
T1 - Knowledge Based Nested Frames of Discernment for Target Integrated Identification
AU - Cui, Yihan
AU - Liang, Yan
AU - Ma, Chaoxiong
AU - Song, Qianqian
AU - Wang, Fan
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - In the context of multiple attributive classification, different frames of discernment (FoD) are often used to address discrepant identity attribute identification for the same target. Though these frames based on different tasks and purposes employ distinct perspectives and structures to identify targets, they exhibit complementary information mapping and transformation modes, revealing nested relationships among multiple FoD. Leveraging this relationship, we can expand the information source of a single FoD and enhance the decision accuracy of integrated target identification. However, the classification results obtained from multiple FoD remain uncertain due to the unknown sensor error in the transformation relation and the randomness of information acquisition. Therefore, this paper aims to develop a integrated identification algorithm for traffic targets based on the transformation relation of nested frames of discernment (NFoD), and construct nested frames transformation matrix and basic belief assignment (BBA) mapping rules using domain rule knowledge and historical data. The nested frame mapping results are obtained by minimizing target distance transformation optimization and calculating the transformation matrix, and the original belief assignment is fused to realize effective transformation between multiple FoD. Simulation results demonstrate that the proposed algorithm can obtain more accurate results compared to the single FoD, which effectively confirms that the transformation fusion method utilizing the NFoD can improve the capability of the integrated target identification system.
AB - In the context of multiple attributive classification, different frames of discernment (FoD) are often used to address discrepant identity attribute identification for the same target. Though these frames based on different tasks and purposes employ distinct perspectives and structures to identify targets, they exhibit complementary information mapping and transformation modes, revealing nested relationships among multiple FoD. Leveraging this relationship, we can expand the information source of a single FoD and enhance the decision accuracy of integrated target identification. However, the classification results obtained from multiple FoD remain uncertain due to the unknown sensor error in the transformation relation and the randomness of information acquisition. Therefore, this paper aims to develop a integrated identification algorithm for traffic targets based on the transformation relation of nested frames of discernment (NFoD), and construct nested frames transformation matrix and basic belief assignment (BBA) mapping rules using domain rule knowledge and historical data. The nested frame mapping results are obtained by minimizing target distance transformation optimization and calculating the transformation matrix, and the original belief assignment is fused to realize effective transformation between multiple FoD. Simulation results demonstrate that the proposed algorithm can obtain more accurate results compared to the single FoD, which effectively confirms that the transformation fusion method utilizing the NFoD can improve the capability of the integrated target identification system.
KW - basic belief assignment
KW - multiple attributive classification
KW - nested frames of discernment
KW - target integrated identification
UR - http://www.scopus.com/inward/record.url?scp=85189349176&partnerID=8YFLogxK
U2 - 10.1109/CAC59555.2023.10450752
DO - 10.1109/CAC59555.2023.10450752
M3 - 会议稿件
AN - SCOPUS:85189349176
T3 - Proceedings - 2023 China Automation Congress, CAC 2023
SP - 1454
EP - 1459
BT - Proceedings - 2023 China Automation Congress, CAC 2023
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2023 China Automation Congress, CAC 2023
Y2 - 17 November 2023 through 19 November 2023
ER -