Search results

1 – 10 of over 1000
Per page
102050
Citations:
Loading...
Access Restricted. View access options
Article
Publication date: 14 December 2023

Huaxiang Song, Chai Wei and Zhou Yong

The paper aims to tackle the classification of Remote Sensing Images (RSIs), which presents a significant challenge for computer algorithms due to the inherent characteristics of…

75

Abstract

Purpose

The paper aims to tackle the classification of Remote Sensing Images (RSIs), which presents a significant challenge for computer algorithms due to the inherent characteristics of clustered ground objects and noisy backgrounds. Recent research typically leverages larger volume models to achieve advanced performance. However, the operating environments of remote sensing commonly cannot provide unconstrained computational and storage resources. It requires lightweight algorithms with exceptional generalization capabilities.

Design/methodology/approach

This study introduces an efficient knowledge distillation (KD) method to build a lightweight yet precise convolutional neural network (CNN) classifier. This method also aims to substantially decrease the training time expenses commonly linked with traditional KD techniques. This approach entails extensive alterations to both the model training framework and the distillation process, each tailored to the unique characteristics of RSIs. In particular, this study establishes a robust ensemble teacher by independently training two CNN models using a customized, efficient training algorithm. Following this, this study modifies a KD loss function to mitigate the suppression of non-target category predictions, which are essential for capturing the inter- and intra-similarity of RSIs.

Findings

This study validated the student model, termed KD-enhanced network (KDE-Net), obtained through the KD process on three benchmark RSI data sets. The KDE-Net surpasses 42 other state-of-the-art methods in the literature published from 2020 to 2023. Compared to the top-ranked method’s performance on the challenging NWPU45 data set, KDE-Net demonstrated a noticeable 0.4% increase in overall accuracy with a significant 88% reduction in parameters. Meanwhile, this study’s reformed KD framework significantly enhances the knowledge transfer speed by at least three times.

Originality/value

This study illustrates that the logit-based KD technique can effectively develop lightweight CNN classifiers for RSI classification without substantial sacrifices in computation and storage costs. Compared to neural architecture search or other methods aiming to provide lightweight solutions, this study’s KDE-Net, based on the inherent characteristics of RSIs, is currently more efficient in constructing accurate yet lightweight classifiers for RSI classification.

Details

International Journal of Web Information Systems, vol. 20 no. 2
Type: Research Article
ISSN: 1744-0084

Keywords

Access Restricted. View access options
Article
Publication date: 26 January 2023

Liang Shao, Liang Wang, Zaiyang Xie and Hua Zhou

Viewing the domestic downside risk as a “pushing” factor for outward foreign direct investment (OFDI), this study aims to examine the surge in Chinese cross-border acquisitions…

175

Abstract

Purpose

Viewing the domestic downside risk as a “pushing” factor for outward foreign direct investment (OFDI), this study aims to examine the surge in Chinese cross-border acquisitions (CBAs) between 2008 and 2017, a unique window when private firms in China were allowed to conduct CBAs.

Design/methodology/approach

This study examines the effect of down-side risk on cross-border acquisition performance by using the sample of Chinese A-share listed companies from 2008 to 2017. Specifically, this study considers three kinds of systemic risk, systematic risk and idiosyncratic risk, and respectively examines their impact on CBAs activities; this study also investigates their subsequent results after CBAs activities. The contingency effect of state ownership on the above relationship is also discussed.

Findings

The findings reveal that pre-CBA systemic risk explains the volume of CBA activities; CBAs are followed by a reduction in systemic risk; the interactions between systemic risk and CBAs decrease with the level of state ownership; and the above results do not hold for traditional risk measures (i.e. systematic risk and idiosyncratic risk).

Originality/value

This study contributes to the literature by revealing the role of systemic risk as a “pushing” factor in the context of OFDI and suggesting an alternative explanation for CBAs from China: Chinese firms (especially private firms) took advantage of the rare opportunity between 2008 and 2017 given by the government to transfer assets overseas through CBA.

Details

Multinational Business Review, vol. 31 no. 3
Type: Research Article
ISSN: 1525-383X

Keywords

Access Restricted. View access options
Article
Publication date: 25 November 2021

Rui Yu and Hua Zhou

Trajectory tracking is an important issue to underactuated unmanned surface vehicles (USVs). However, parametric uncertainties and environmental disturbances bring great…

163

Abstract

Purpose

Trajectory tracking is an important issue to underactuated unmanned surface vehicles (USVs). However, parametric uncertainties and environmental disturbances bring great challenges to the precise trajectory tracking control of USVs. This paper aims to propose a robust trajectory tracking control algorithm with exponential stability for underactuated USVs with parametric uncertainties and unknown environmental disturbances.

Design/methodology/approach

In this method, the backstepping method and sliding mode control method are combined to ensure that the underactuated USV can track and maintain the desired trajectory. In addition, a modified switching-gain adaptation algorithm is adopted to enhance the robustness and reduce chattering. Besides, the global exponential stability of the closed-loop system is proved by Lyapunov’s direct method.

Findings

The proposed method in this paper offers a robust trajectory tracking solution to underactuated USVs and it is verified by simulations and experiments. Compared with the traditional proportion-integral-derivative method and several state-of-the-art algorithms, the proposed method has superior performance in simulation and experimental results.

Originality/value

This paper proposes a robust trajectory tracking control algorithm with exponential stability for underactuated USVs. The proposed method achieves exponential stability with better robustness and transient performance.

Details

Industrial Robot: the international journal of robotics research and application, vol. 49 no. 4
Type: Research Article
ISSN: 0143-991X

Keywords

Access Restricted. View access options
Article
Publication date: 29 January 2020

Fei Zhang, Xiao-Hua Zhou, Jiafu Su, Sang-Bing Tsai and Yu-Ming Zhai

The purpose of this paper is to examine how signals of uncertainty in the media affect retail investor decisions and initial public offering (IPO) underpricing through theoretical…

380

Abstract

Purpose

The purpose of this paper is to examine how signals of uncertainty in the media affect retail investor decisions and initial public offering (IPO) underpricing through theoretical and empirical methods.

Design/methodology/approach

The authors construct a theoretical model of the influence of media signals on IPO pricing, which describes the micro process in which uncertain signals in media influence retail investors’ decisions and IPO underpricing. Besides, the authors take 516 small and medium-size enterprises (SMEs) listed in A-share from July 2009 to December 2012 as samples for empirical tests and establish an in-depth learning model for text analysis with Java programming to measure Chinese media tone. Finally, the results of the model analysis are verified by empirical results.

Findings

The results show that authoritative media with high credibility can reduce the uncertainty of information sources attract more investors’ attention and improve the valuation and demand of retail investors. The higher the media credibility is the higher the IPO underpricing rate is. The uncertain tone of the media will increase the decision-making cost of investors reduce the valuation expectation and demand of the secondary market and lead to a lower IPO underpricing rate.

Originality/value

The authors study the influence of the uncertainty of media source and media content on the degree of IPO underpricing of SMEs. This is a useful supplement to the Chinese media tone research system that is still in the exploration stage. The research has reference value for government regulation and investor decision-making.

Access Restricted. View access options
Article
Publication date: 18 January 2021

Hua Zhou, Dong Wei, Yinglong Chen and Fa Wu

To promote the intuitiveness of collaborative tasks, the negotiation ability of humans with each other has inspired a large amount of studies aimed at reproducing the capacity in…

233

Abstract

Purpose

To promote the intuitiveness of collaborative tasks, the negotiation ability of humans with each other has inspired a large amount of studies aimed at reproducing the capacity in physical human-robot interaction (pHRI). This paper aims to promote mutual adaptation in negotiation when both parties possess incomplete information.

Design/methodology/approach

This paper introduces virtual fixtures into the traditional negotiation mechanism, locally regulating tracking trajectory and impedance parameters in the negotiating phase until the final plan integrates bilateral intentions well. In the strategy, robots convey its task information to humans and offer groups of guide plans for them to choose, on the premise of maximizing the robot’s own profits.

Findings

Compared with traditional negotiation strategies, humans adapt to robots easily and show lower cognitive load in the method, while the satisfied plan shows better performance for the whole human-robot system.

Originality/value

In this study, this paper proposes a novel negotiation strategy to facilitate the mutual adaptation of humans and robots in complicated shared tasks, especially when both parties possess incomplete information of tasks.

Details

Industrial Robot: the international journal of robotics research and application, vol. 48 no. 2
Type: Research Article
ISSN: 0143-991X

Keywords

Access Restricted. View access options
Article
Publication date: 9 April 2018

Ruilong Du, Yinglong Chen and Hua Zhou

The purpose of this paper is to propose a simulation model for studying the lubricating gap between the ring gear and the case in internal gear pumps.

229

Abstract

Purpose

The purpose of this paper is to propose a simulation model for studying the lubricating gap between the ring gear and the case in internal gear pumps.

Design/methodology/approach

The pressure distribution of the wedge-shaped oil film between the ring gear and the case is obtained based on the theory of film lubrication using the Reynolds equation implemented with MATLAB. After that, the balance of the ring gear is achieved by the radial micro motion of the ring gear. The power loss due to the leakage and the shear stress is then calculated for optimized design of the radial clearance.

Findings

The hydrodynamic effect and the squeezing effect of the wedge-shaped oil film play a role in the hydrodynamic balance of the ring gear, and they become more intense when the operating speed gets lower and the pressure gets higher. The optimal radial clearance should stay between 20 and 25 µm for the minimum power loss.

Originality/value

The present research provides the first simulation model that treats the oil film between the ring gear and the case as wedge-shaped oil film and explains why the ring gear stays balanced. Furthermore, the simulation model can be regarded as a tool for optimized design of the radial clearance.

Details

Industrial Lubrication and Tribology, vol. 70 no. 3
Type: Research Article
ISSN: 0036-8792

Keywords

Access Restricted. View access options
Article
Publication date: 8 March 2022

Rui Yu, Hua Zhou, Siyu Ma, Guifu Luo and Mingwei Lin

Hydrodynamic parameter estimation is significant for the velocity prediction of unmanned surface vehicles. Considering the field data’s uncertain nonlinearities (environmental…

94

Abstract

Purpose

Hydrodynamic parameter estimation is significant for the velocity prediction of unmanned surface vehicles. Considering the field data’s uncertain nonlinearities (environmental disturbances and measurement noise), this paper aims to propose a hybrid adaptive parameter estimation (HAPE) strategy.

Design/methodology/approach

First, a rough estimation of hydrodynamic parameters is used by the least squares method. Second, an improved adaptive parameter estimation algorithm is applied to compensate for the influence of uncertain nonlinearities and adjust the parameters within the rough range. Finally, it is proved that the calculated velocity asymptotically converges to the actual value during the parameter estimation procedure.

Findings

The numerical simulation and pool experiments are conducted in two scenarios of steady turning and sinusoidal thrust to verify the effectiveness of the proposed HAPE method. The results validate that the accuracy of the predicted velocity using the hydrodynamic model obtained by the HAPE strategy is better than the APE algorithm. In addition, the hydrodynamic parameters estimated with the sinusoidal thrust data are more applicable than the steady turning data.

Originality/value

This study proposes a HAPE strategy that considers the uncertain nonlinearities of the field data. This method provides a more accurate predicted velocity. Besides, as far as we know, it is the first time to analyze the influence of different test conditions on the accuracy of the predicted velocity.

Details

Industrial Robot: the international journal of robotics research and application, vol. 49 no. 4
Type: Research Article
ISSN: 0143-991X

Keywords

Access Restricted. View access options
Article
Publication date: 10 May 2024

Lu Li and Dong-hua Zhou

This paper aims to obtain a calculation method by hand without iteration.

37

Abstract

Purpose

This paper aims to obtain a calculation method by hand without iteration.

Design/methodology/approach

This paper adopts strains as known quantities to solve the internal forces and deformations of the section, simplifies the deflection curve of the column and obtains nomograms that can calculate the bearing capacity and reinforcement of circular reinforced concrete (RC) columns by hand.

Findings

Nomograms include five variables: mechanical reinforcement ratio, relative normal force, dimensionless bending moment, slenderness ratio and ultimate dimensionless curvature. Nomograms corresponding to all classes of concrete have been drawn, and their dimensionless form makes them widely applicable. The calculation results of nomograms are compared and analysed with numerical calculation results, and the difference is within 5%, meeting the engineering requirements.

Originality/value

Calculating the bearing capacity of compression bending components requires considering second-order effects. Therefore, the calculation of the bearing capacity of circular RC columns requires iterative calculation, as it includes dual nonlinearity of material and geometry, and the two are coupled with each other. To calculate the bearing capacity of the section adopting ordinary concrete, it is necessary to solve the transcendental equation iteratively. For high-strength concrete, it can only be solved by numerical integration. A fast calculation method by hand is proposed in this paper.

Details

International Journal of Structural Integrity, vol. 15 no. 4
Type: Research Article
ISSN: 1757-9864

Keywords

Access Restricted. View access options
Article
Publication date: 13 November 2024

Huaxiang Song, Hanjun Xia, Wenhui Wang, Yang Zhou, Wanbo Liu, Qun Liu and Jinling Liu

Vision transformers (ViT) detectors excel in processing natural images. However, when processing remote sensing images (RSIs), ViT methods generally exhibit inferior accuracy…

13

Abstract

Purpose

Vision transformers (ViT) detectors excel in processing natural images. However, when processing remote sensing images (RSIs), ViT methods generally exhibit inferior accuracy compared to approaches based on convolutional neural networks (CNNs). Recently, researchers have proposed various structural optimization strategies to enhance the performance of ViT detectors, but the progress has been insignificant. We contend that the frequent scarcity of RSI samples is the primary cause of this problem, and model modifications alone cannot solve it.

Design/methodology/approach

To address this, we introduce a faster RCNN-based approach, termed QAGA-Net, which significantly enhances the performance of ViT detectors in RSI recognition. Initially, we propose a novel quantitative augmentation learning (QAL) strategy to address the sparse data distribution in RSIs. This strategy is integrated as the QAL module, a plug-and-play component active exclusively during the model’s training phase. Subsequently, we enhanced the feature pyramid network (FPN) by introducing two efficient modules: a global attention (GA) module to model long-range feature dependencies and enhance multi-scale information fusion, and an efficient pooling (EP) module to optimize the model’s capability to understand both high and low frequency information. Importantly, QAGA-Net has a compact model size and achieves a balance between computational efficiency and accuracy.

Findings

We verified the performance of QAGA-Net by using two different efficient ViT models as the detector’s backbone. Extensive experiments on the NWPU-10 and DIOR20 datasets demonstrate that QAGA-Net achieves superior accuracy compared to 23 other ViT or CNN methods in the literature. Specifically, QAGA-Net shows an increase in mAP by 2.1% or 2.6% on the challenging DIOR20 dataset when compared to the top-ranked CNN or ViT detectors, respectively.

Originality/value

This paper highlights the impact of sparse data distribution on ViT detection performance. To address this, we introduce a fundamentally data-driven approach: the QAL module. Additionally, we introduced two efficient modules to enhance the performance of FPN. More importantly, our strategy has the potential to collaborate with other ViT detectors, as the proposed method does not require any structural modifications to the ViT backbone.

Details

International Journal of Intelligent Computing and Cybernetics, vol. ahead-of-print no. ahead-of-print
Type: Research Article
ISSN: 1756-378X

Keywords

Access Restricted. View access options
Article
Publication date: 22 February 2021

Bai Yun, Zhao Yue and Zhou Yaolin

This study aims to identify the prominent topics, the distribution and association characteristics of topics and the topic evolutionary trends of Documentary Heritage Preservation…

995

Abstract

Purpose

This study aims to identify the prominent topics, the distribution and association characteristics of topics and the topic evolutionary trends of Documentary Heritage Preservation and Conservation (DHPAC) research in China.

Design/methodology/approach

Keywords of relevant papers in China National Knowledge Infrastructure (CNKI) were extracted as the data source in this study. First, frequency and co-occurrence of keywords of the selected papers were obtained by using SATI. Second, co-word network indicators were calculated with the Pajek software. Then, VOSviewer was applied to optimize the visualization of the sub-communities. Finally, a topics evolution map of this research field was implemented by CorTexT.

Findings

The research topics of DHPAC research in China were unbalanced but distinct. Topics of DHPAC research in China possessed inconspicuous orientation and consistency. The core topics had less influence on the overall network. A research system had formed with archival conservation and ancient books conservation as the core research directions. Research in this field had formed four continuous evolutionary paths about ancient books conservation, salvage conservation, archival conservation and archives conservation technology science with topics fusion and differentiation coexisting. Attentions on “ancient books conservation”, “paper relics conservation”, “electronic record”, “digitization”, “minority”, “documents in the republic of China” had increased during the past two decades and new hot topics of DHPAC research kept appearing in China.

Originality/value

This study synthesized and analyzed the research results of DHPAC research in China from a more comprehensive perspective and revealed the topic structure and longitudinal evolution process intuitively with co-word analysis and social network analysis, which can assist researchers to improve research systematization, discover new research directions and seek cooperative research path.

1 – 10 of over 1000
Per page
102050