Bo Wu

Bo Wu

Auto-generated profile, based on publicly available information

Bo Wu

1 Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai, China

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 November 2022 in IEEE Transactions on Industrial Informatics
Verified
Journal Article
Closed access
YZYang Zhang
Yang Zhang
YZYang Zhou
Yang Zhou
HPHuilin Pan
Huilin Pan
BWBo Wu
Bo Wu
GSGuodong Sun
Guodong Sun
Fault detection for key components in the braking system of freight trains is critical for ensuring railway transportation safety. Despite the frequently employed methods based on deep learning, these fault detectors are extremely reliant on hardware resources and complex to implement. In addition, no train fault detectors consider the drop in accuracy induced by scale variation of fault parts. This article proposes a lightweight anchor-free framework to solve the above problems. Specifically, to reduce the amount of computation and model size, we introduce a lightweight backbone and adopt an anchor-free method for localization and regression. To improve detection accuracy for multiscale parts, we design a feature pyramid network to generate rectangular layers of different sizes to map parts with similar aspect ratios. Experiments on four fault datasets show that our framework achieves 98.44% accuracy while the model size is only 22.5 MB, outperforming the state-of-the-art detectors.

10 April 2024 in IEEE Transactions on Industrial Informatics
Verified
Journal Article
Closed access
GSGuodong Sun
Guodong Sun
YPYuting Peng
Yuting Peng
LCLe Cheng
Le Cheng
MXMengya Xu
Mengya Xu
AWAn Wang
An Wang
YZYang Zhang
Yang Zhang
The precise segmentation of ore images is critical to the successful execution of the beneficiation process. Due to the homogeneous appearance of the ores, which leads to low contrast and unclear boundaries, accurate segmentation becomes challenging, and recognition becomes problematic. This article proposes a lightweight framework based on multilayer perceptron (MLP), which focuses on solving the problem of edge blurring. Specifically, we introduce a lightweight backbone better suited for efficiently extracting low-level features. Besides, we design a feature pyramid network consisting of two MLP structures that balance local and global information, thus, enhancing detection accuracy. Furthermore, we propose a novel loss function that guides the prediction points to match the instance edge points to achieve clear object boundaries. We have conducted extensive experiments to validate the efficacy of our proposed method. Our approach achieves a remarkable processing speed of over 27 frames per second with a model size of only 73 MB. Moreover, our method delivers a consistently high level of accuracy, with impressive performance scores of 60.4 and 48.9 in $AP_{50}^{\text{box}}$ and $AP_{50}^{\text{mask}}$ , respectively, as compared with the currently available state-of-the-art techniques, when tested on the ORE image dataset.

29 September 2018 in Entropy
Verified
Journal Article
Open access
PDF file
BWBo Wu
Bo Wu
YGYangde Gao
Yangde Gao
SFSonglin Feng
Songlin Feng
TCTheerasak Chanwimalueang
Theerasak Chanwimalueang
To reduce the maintenance cost and safeguard machinery operation, remaining useful life (RUL) prediction is very important for long term health monitoring. In this paper, we introduce a novel hybrid method to deal with the RUL prediction for health management. Firstly, the sparse reconstruction algorithm of the optimized Lasso and the Least Square QR-factorization (Lasso-LSQR) is applied to compressed sensing (CS), which can realize the sparse optimization for long term health monitoring data. After the sparse signal is reconstructed, the minimum entropy de-convolution (MED) is used to identify the fault characteristics and to obtain significant fault information from the machinery operation. Health indicators with Skip-over, sample entropy and approximate entropy are then performed to track the degradation of the machinery process. The performance analysis of the Skip-over is superior to other indicators. Finally, Fractal Autoregressive Integrated Moving Average model (FARIMA) is employed to predict the Skip-over using the R/S method. The analysis results evidence that the novel hybrid method yields a good performance, and such method can achieve highly accurate RUL prediction and safeguard machinery operation for long term monitoring.

23 May 2022 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
GSGuodong Sun
Guodong Sun
YZYang Zhou
Yang Zhou
HPHuilin Pan
Huilin Pan
BWBo Wu
Bo Wu
YHYe Hu
Ye Hu
YZYang Zhang
Yang Zhang
Real-time vision-based system of fault detection (RVBS-FD) for freight trains is an essential part of ensuring railway transportation safety. Most existing vision-based methods still have high computational costs based on convolutional neural networks. The computational cost is mainly reflected in the backbone, neck, and postprocessing, i.e., nonmaximum suppression (NMS). In this article, we propose a lightweight NMS-free framework to achieve real-time detection and high accuracy simultaneously. First, we use a lightweight backbone for feature extraction and design a fault detection pyramid to process features. This fault detection pyramid includes three novel individual modules using attention mechanism, bottleneck, and dilated convolution for feature enhancement and computation reduction. Instead of using NMS, we calculate different loss functions, including classification and location costs in the detection head, to further reduce computation. Experimental results show that our framework achieves over 83-frames/s speed with a smaller model size and higher accuracy than the state-of-the-art detectors. Meanwhile, the hardware resource requirements of our method are low during the training and testing process.

1 January 2021 in Shock and Vibration
Verified
Journal Article
Open access
PDF file
GSGuodong Sun
Guodong Sun
YHYe Hu
Ye Hu
BWBo Wu
Bo Wu
HZHongyu Zhou
Hongyu Zhou
Addressing the problem that it is difficult to extract the features of vibration signal and diagnose the fault of rolling bearing, we propose a novel diagnosis method combining multisynchrosqueezing S transform and faster dictionary learning (MSSST‐FDL). Firstly, MSSST is adopted to transform vibration signals into high‐resolution time‐frequency images. Then, the local binary pattern (LBP) operator is introduced to extract the low‐dimensional texture features of time‐frequency images, which improves the speed of fault recognition. Finally, nonnegative matrix factorization (NMF) with only one hyperparameter and nonnegative linear equation are used to solve the dictionary learning and feature coding, respectively. The feature coding is input into the classifier for training and recognition. Experiments show that our method performs well on the rolling bearing dataset of Case Western Reserve University (CWRU) and the Society for Machinery Failure Prevention Technology (MFPT). Further, the proposed method is applied to the loudspeaker pure‐tone detection dataset, and the loudspeaker anomaly diagnosis is achieved. The diagnosis results verify that our method can meet the needs of practical engineering.

16 July 2025 in Knowledge-Based Systems
Verified
Journal Article
Closed access
GSGuodong Sun
Guodong Sun
JLJunjie Liu
Junjie Liu
GZGaoyang Zhang
Gaoyang Zhang
BWBo Wu
Bo Wu
YZYang Zhang
Yang Zhang