Hongbo Fan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hongbo Fan has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Qingzhu Li
Department of Vehicle and Electrical Engineering, Army Engineering University of PLA, Shijiazhuang, China
2 shared publications - Zhining Li
Department of Vehicle and Electrical Engineering, Army Engineering University of PLA, Shijiazhuang, China
2 shared publications - Zhiyong Shi
Department of Vehicle and Electrical Engineering, Army Engineering University of PLA, Shijiazhuang, China
2 shared publications - Dinghai Wu
Army Engineering University of PLA, Shijiazhuang Campus, Shijiazhuang, China
1 shared publication - Guang Zhang
Army Academy of Artillery and Air Defense of PLA, Shenyang, China
1 shared publication Ren Guoquan
Shijiazhuang Campus, Army Engineering University of PLA, Shijiazhuang, China
1 shared publication
Latest Publications
- 17 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-15
Multi-Target Magnetic Positioning Using SAFCM Clustering and Invariants-Improved Tilt Angle
The positioning of multiple magnetic sources with different depths, distributions and magnetic moments is extremely difficult. The improved tilt angle for target area recognition can be well adapted to the self-adaptive fuzzy c-means (SAFCM) clustering algorithm for grid measurement of magnetic gradient tensors (MGT), which provides a feasible solution to solve the complex problem of multi-target inversion. First, we improve the tilt angle with the normalized source strengths and the tensor contraction, which can accurately identify the two-dimensional area boundary of objects with different buried depths regardless of the magnetization direction. The grid measurements of magnetic gradient tensor systems (MGTS) enable this process. Then, with the tensor spacial invariant relations, we calculate the source initial position coordinates corresponding to each node in the recognition area, thereby forming a dense point cloud in the space near the target real position. Finally, we introduce a penalty function into the clustering validity function to eliminate the monotonous decreasing trend of the cluster number optimization process, thus the SAFCM algorithm can accurately and automatically detect the number and position of centroids. The simulation shows that with the Gaussian noise of 0 mean and 5 nT/m variance, the targets number recognition accuracy of 20 magnetic dipoles is 97.6%, and the horizontal position and buried depth estimation accuracy is greater than 91.7% and 85.6% respectively. In the real survey areas of 2.1m×2.1m and 1.2m×1.2m, the coordinates’ estimation deviation of nine small magnets is less than 0.091m.
- 17 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-15
Multi-Target Magnetic Positioning Using SAFCM Clustering and Invariants-Improved Tilt Angle
The positioning of multiple magnetic sources with different depths, distributions and magnetic moments is extremely difficult. The improved tilt angle for target area recognition can be well adapted to the self-adaptive fuzzy c-means (SAFCM) clustering algorithm for grid measurement of magnetic gradient tensors (MGT), which provides a feasible solution to solve the complex problem of multi-target inversion. First, we improve the tilt angle with the normalized source strengths and the tensor contraction, which can accurately identify the two-dimensional area boundary of objects with different buried depths regardless of the magnetization direction. The grid measurements of magnetic gradient tensor systems (MGTS) enable this process. Then, with the tensor spacial invariant relations, we calculate the source initial position coordinates corresponding to each node in the recognition area, thereby forming a dense point cloud in the space near the target real position. Finally, we introduce a penalty function into the clustering validity function to eliminate the monotonous decreasing trend of the cluster number optimization process, thus the SAFCM algorithm can accurately and automatically detect the number and position of centroids. The simulation shows that with the Gaussian noise of 0 mean and 5 nT/m variance, the targets number recognition accuracy of 20 magnetic dipoles is 97.6%, and the horizontal position and buried depth estimation accuracy is greater than 91.7% and 85.6% respectively. In the real survey areas of 2.1m×2.1m and 1.2m×1.2m, the coordinates’ estimation deviation of nine small magnets is less than 0.091m.
- 14 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 70 (00189456) , 1-14
Preferred Configuration and Detection Limits Estimation of Magnetic Gradient Tensor System
The magnetic gradient tensor system (MGTS) based on the fluxgate sensors array replaces partial differentials with the sensors reading difference to approximate the magnetic field gradient, but it introduces structural errors at the mean time. A good structure design with reasonable parameters can reduce structural errors and expand the theoretical detection range of the MGTS. From the analysis of the complete 18-uniaxial MGTS, five simplified structures of the MGTS were summarized, which include the plane cross, equilateral triangle, square, right-angle tetrahedron, and regular tetrahedron. The differential magnetic gradient tensor measurement matrixes of different structures were also provided, where we found the plane-cross structure produces the smallest structural error. Moreover, the theoretical detection limits expression of the MGTS to the magnetic dipole was derived. The detection limit boundary of the MGTS is related to the sensor’s accuracy, the array baseline distance, and the magnetization direction and magnetic moment magnitude of the source. In actual heading-line surveys, the magnetic moments of four typical small magnets were successfully estimated by the magnetic dipole positioning method, and the theoretical detection limits of the built plane-cross MGTS aiming at the four magnets were estimated. The results show that the estimation accuracy of the theoretical detection limits of MGTs is ±0.5 m.
- 14 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 70 (00189456) , 1-14
Preferred Configuration and Detection Limits Estimation of Magnetic Gradient Tensor System
The magnetic gradient tensor system (MGTS) based on the fluxgate sensors array replaces partial differentials with the sensors reading difference to approximate the magnetic field gradient, but it introduces structural errors at the mean time. A good structure design with reasonable parameters can reduce structural errors and expand the theoretical detection range of the MGTS. From the analysis of the complete 18-uniaxial MGTS, five simplified structures of the MGTS were summarized, which include the plane cross, equilateral triangle, square, right-angle tetrahedron, and regular tetrahedron. The differential magnetic gradient tensor measurement matrixes of different structures were also provided, where we found the plane-cross structure produces the smallest structural error. Moreover, the theoretical detection limits expression of the MGTS to the magnetic dipole was derived. The detection limit boundary of the MGTS is related to the sensor’s accuracy, the array baseline distance, and the magnetization direction and magnetic moment magnitude of the source. In actual heading-line surveys, the magnetic moments of four typical small magnets were successfully estimated by the magnetic dipole positioning method, and the theoretical detection limits of the built plane-cross MGTS aiming at the four magnets were estimated. The results show that the estimation accuracy of the theoretical detection limits of MGTs is ±0.5 m.
- 23 October 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 877-882
Mechanical Fault Diagnosis based on Dual-tree Complex Wavelet Packet Time-frequency Distribution and Residual Network Transfer Learning
Aiming at the problem of large amount of calculation for the time-frequency distribution of mechanical vibration signals with good representation ability, a mechanical fault diagnosis method based on dual-tree complex wavelet packet time-frequency distribution and residual network transfer learning is proposed. Firstly, the high-frequency and low-frequency components of the signal are decomposed based on the dual-tree complex wavelet packet with translation invariance, and the discrete time-frequency distribution of the mechanical vibration signal is efficiently obtained. In view of the fact that ResNet uses residual modules and cross-layer connections to build a deep network structure, it has strong image recognition and classification capabilities, Residual network transfer learning is used to train and classify the obtained time-frequency distribution images, and the Western Reserve University bearing dataset and the engines measured signal are used to verify the method proposed in this paper, and it can obtain ideal fault diagnosis results under different input images sizes and a small number of training sample conditions.
- 23 October 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 877-882
Mechanical Fault Diagnosis based on Dual-tree Complex Wavelet Packet Time-frequency Distribution and Residual Network Transfer Learning
Aiming at the problem of large amount of calculation for the time-frequency distribution of mechanical vibration signals with good representation ability, a mechanical fault diagnosis method based on dual-tree complex wavelet packet time-frequency distribution and residual network transfer learning is proposed. Firstly, the high-frequency and low-frequency components of the signal are decomposed based on the dual-tree complex wavelet packet with translation invariance, and the discrete time-frequency distribution of the mechanical vibration signal is efficiently obtained. In view of the fact that ResNet uses residual modules and cross-layer connections to build a deep network structure, it has strong image recognition and classification capabilities, Residual network transfer learning is used to train and classify the obtained time-frequency distribution images, and the Western Reserve University bearing dataset and the engines measured signal are used to verify the method proposed in this paper, and it can obtain ideal fault diagnosis results under different input images sizes and a small number of training sample conditions.