Sendren Sheng-Dong Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Sendren Sheng-Dong Xu has a total of 81 co-authors. Here are the top co-authors by number of shared publications:
- Hsu-Chih Huang
Department of Mechanical Engineering and Advanced Institute of Manufacturing With High‐Tech Innovations National Chung Cheng University Chiayi Taiwan
9 shared publications Getachew Nadew Wedajew
Department of Power Mechanical Engineering, National Tsing Hua University, Hsinchu, Taiwan
3 shared publications- Getachew Nadew Wedajew
Department of Power Mechanical Engineering National Tsing Hua University Hsinchu Taiwan
3 shared publications Sharfiden Hassen Yusuf
College of Engineering and Physics, King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
3 shared publications- Chien-Tien Su
Department of Family Medicine, Taipei Medical University Hospital, Taipei 11031, Taiwan
2 shared publications Chih-Hao Chiang
National Ilan University
2 shared publications
Latest Publications
- 10 November 2026
- Published by Wiley in Asian Journal of Control
Evolutionary Fuzzy Optimization Control for Computational Intelligence Based Robust Visual Simultaneous Localization and Mapping of Autonomous Mobile Robots
This study presents a contribution to the development of evolutionary fuzzy optimization control for computational intelligence (CI)‐based robust visual simultaneous localization and mapping (VSLAM) in autonomous mobile robots. The proposed SLAM method integrates visual perception, image processing, model predictive control (MPC), and fuzzy logic to improve localization robustness and map consistency. Some CI techniques are employed for image processing in the proposed ORB (Oriented fast and Rotated Brief)‐SLAM3 VSLAM framework. An interval Type‐2 fuzzy system is incorporated with MPC to develop the intelligent SLAM controller, called IT2FMPC for autonomous mobile robots. Moreover, an evolutionary genetic algorithm (GA) optimization strategy is employed to adaptively refine fuzzy membership functions and rule sets. The optimized GA‐IT2FMPC controller is integrated into the VSLAM to improve feature selection, motion estimation, and loop‐closure consistency. Numerical simulations and comparative studies conducted on publicly available datasets validate the performance and superiority of the developed VSLAM method. A physical experimental differential‐drive robot is constructed to demonstrate that the proposed framework significantly reduces pose error and generates more consistent maps when compared with representative state‐of‐the‐art methods.
- 10 November 2026
- Published by Wiley in Asian Journal of Control
Evolutionary Fuzzy Optimization Control for Computational Intelligence Based Robust Visual Simultaneous Localization and Mapping of Autonomous Mobile Robots
This study presents a contribution to the development of evolutionary fuzzy optimization control for computational intelligence (CI)‐based robust visual simultaneous localization and mapping (VSLAM) in autonomous mobile robots. The proposed SLAM method integrates visual perception, image processing, model predictive control (MPC), and fuzzy logic to improve localization robustness and map consistency. Some CI techniques are employed for image processing in the proposed ORB (Oriented fast and Rotated Brief)‐SLAM3 VSLAM framework. An interval Type‐2 fuzzy system is incorporated with MPC to develop the intelligent SLAM controller, called IT2FMPC for autonomous mobile robots. Moreover, an evolutionary genetic algorithm (GA) optimization strategy is employed to adaptively refine fuzzy membership functions and rule sets. The optimized GA‐IT2FMPC controller is integrated into the VSLAM to improve feature selection, motion estimation, and loop‐closure consistency. Numerical simulations and comparative studies conducted on publicly available datasets validate the performance and superiority of the developed VSLAM method. A physical experimental differential‐drive robot is constructed to demonstrate that the proposed framework significantly reduces pose error and generates more consistent maps when compared with representative state‐of‐the‐art methods.
- 3 October 2026
- Published by Elsevier in Digital Signal Processing
- Vol. 185
Eliminating rainy night image blurs caused by raindrops adhering to the lens based on end-to-end residual neural networks
- 3 October 2026
- Published by Elsevier in Digital Signal Processing
- Vol. 185
Eliminating rainy night image blurs caused by raindrops adhering to the lens based on end-to-end residual neural networks
- 12 May 2026
- Published by Elsevier in Digital Signal Processing
- Vol. 181
An improved sparse R-CNN approach for strawberry disease detection
- 12 May 2026
- Published by Elsevier in Digital Signal Processing
- Vol. 181
An improved sparse R-CNN approach for strawberry disease detection
- 3 February 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (2) , 104
Cross-modal spatio-temporal fusion weakly supervised video anomaly detection based on large-scale vision-language models
- 3 February 2026
- Published by Springer Nature in Multimedia Systems
- Vol. 32 (2) , 104
Cross-modal spatio-temporal fusion weakly supervised video anomaly detection based on large-scale vision-language models
- 10 April 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 163
Convolutional neural network-based single-image contrast enhancement using multi-exposure training
- 10 April 2025
- Published by Elsevier in Digital Signal Processing
- Vol. 163
Convolutional neural network-based single-image contrast enhancement using multi-exposure training
- 10 March 2025
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 17 (4) , 743-761
Human Behavioral, Subjective, and Physiological Assessments Under an Oral Test With a Humanoid Robot Examiner
- 10 March 2025
- Published by Springer Nature in International Journal of Social Robotics
- Vol. 17 (4) , 743-761
Human Behavioral, Subjective, and Physiological Assessments Under an Oral Test With a Humanoid Robot Examiner
- 5 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (4) , 5914-5927
Design and Implementation of a Multi-Threading Architecture for Downlink Process in Software- Defined Physical Layer Over Long-Term Evolution Networks
To engage with new and ongoing trends regarding the technical characteristics and potential uses of the 5th Generation Mobile Networks (5G), EURECOM (a leading European Graduate School and Research Center in the field of digital science in Sophia Antipolis, France) has been devoted to the realization of softwarization of the communication base station (BS) and has developed the open-source code OpenAirInterface. Such realization of the Evolved Node B (eNB) with Software-Defined Radio Access Network (SD-RAN) can transcend the bottlenecks of existing BS architecture and effectively increase transmission efficiency. Here, we analyzed and collated the physical layer (PHY) of the original downlink process in OpenAirInterface (OAI), and then reconstructed the software architecture to improve the computing efficiency for multiple users. The functionality of POSIX Threads (Pthreads) leads to a refreshed process that is packaged into sub-functions, and then passed into a number of newly opened threads for execution in order to i) establish the multi-threading PHY with parallel processing in OAI Downlink; ii) maximize the efficiency of the utilization of hardware resources; iii) reduce the execution time of the downlink process; and iv) expand the capacity of the user equipment (UE) connection and improve connection throughput. We physically used the OAI platform to implement a Long-Term Evolution (LTE) 3GPP-based softwarized BS test. Finally, real experiments show that the proposed approach can successfully parallelize the execution of the downlink process in OAI. This process effectively reduced execution time and offered broad scalability.
- 5 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (4) , 5914-5927
Design and Implementation of a Multi-Threading Architecture for Downlink Process in Software- Defined Physical Layer Over Long-Term Evolution Networks
To engage with new and ongoing trends regarding the technical characteristics and potential uses of the 5th Generation Mobile Networks (5G), EURECOM (a leading European Graduate School and Research Center in the field of digital science in Sophia Antipolis, France) has been devoted to the realization of softwarization of the communication base station (BS) and has developed the open-source code OpenAirInterface. Such realization of the Evolved Node B (eNB) with Software-Defined Radio Access Network (SD-RAN) can transcend the bottlenecks of existing BS architecture and effectively increase transmission efficiency. Here, we analyzed and collated the physical layer (PHY) of the original downlink process in OpenAirInterface (OAI), and then reconstructed the software architecture to improve the computing efficiency for multiple users. The functionality of POSIX Threads (Pthreads) leads to a refreshed process that is packaged into sub-functions, and then passed into a number of newly opened threads for execution in order to i) establish the multi-threading PHY with parallel processing in OAI Downlink; ii) maximize the efficiency of the utilization of hardware resources; iii) reduce the execution time of the downlink process; and iv) expand the capacity of the user equipment (UE) connection and improve connection throughput. We physically used the OAI platform to implement a Long-Term Evolution (LTE) 3GPP-based softwarized BS test. Finally, real experiments show that the proposed approach can successfully parallelize the execution of the downlink process in OAI. This process effectively reduced execution time and offered broad scalability.
- Open Access
- 19 July 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (14) , 2642
SE-RRACycleGAN: Unsupervised Single-Image Deraining Using Squeeze-and-Excitation-Based Recurrent Rain-Attentive CycleGAN
In computer vision tasks, the ability to remove rain from a single image is a crucial element to enhance the effectiveness of subsequent high-level tasks in rainy conditions. Recently, numerous data-driven single-image deraining techniques have emerged, primarily relying on paired images (i.e., in a supervised manner). However, when dealing with real deraining tasks, it is common to encounter unpaired images. In such scenarios, removing rain streaks in an unsupervised manner becomes a challenging task, as there are no constraints between images, resulting in suboptimal restoration results. In this paper, we introduce a new unsupervised single-image deraining method called SE-RRACycleGAN, which does not require a paired dataset for training and can effectively leverage the constrained transfer learning capability and cyclic structures inherent in CycleGAN. Since rain removal is closely associated with the analysis of texture features in an input image, we proposed a novel recurrent rain attentive module (RRAM) to enhance rain-related information detection by simultaneously considering both rainy and rain-free images. We also utilize the squeeze-and-excitation enhancement technique to the generator network to effectively capture spatial contextual information among channels. Finally, content loss is introduced to enhance the visual similarity between the input and generated images. Our method excels at removing numerous rain streaks, preserving a smooth background, and closely resembling the ground truth compared to other approaches, based on both quantitative and qualitative results, without the need for paired training images. Extensive experiments on synthetic and real-world datasets demonstrate that our approach shows superiority over most unsupervised state-of-the-art techniques, particularly on the Rain12 dataset (achieving a PSNR of 34.60 and an SSIM of 0.954) and real rainy images (achieving a PSNR of 34.17 and an SSIM of 0.953), and is highly competitive when compared to supervised methods. Moreover, the performance of our model is evaluated using RMSE, FSIM, MAE, and the correlation coefficient, achieving remarkable results that indicate a high degree of accuracy in rain removal and strong preservation of the original image’s structural details.
- 19 July 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (14) , 2642
Open AccessSE-RRACycleGAN: Unsupervised Single-Image Deraining Using Squeeze-and-Excitation-Based Recurrent Rain-Attentive CycleGAN
In computer vision tasks, the ability to remove rain from a single image is a crucial element to enhance the effectiveness of subsequent high-level tasks in rainy conditions. Recently, numerous data-driven single-image deraining techniques have emerged, primarily relying on paired images (i.e., in a supervised manner). However, when dealing with real deraining tasks, it is common to encounter unpaired images. In such scenarios, removing rain streaks in an unsupervised manner becomes a challenging task, as there are no constraints between images, resulting in suboptimal restoration results. In this paper, we introduce a new unsupervised single-image deraining method called SE-RRACycleGAN, which does not require a paired dataset for training and can effectively leverage the constrained transfer learning capability and cyclic structures inherent in CycleGAN. Since rain removal is closely associated with the analysis of texture features in an input image, we proposed a novel recurrent rain attentive module (RRAM) to enhance rain-related information detection by simultaneously considering both rainy and rain-free images. We also utilize the squeeze-and-excitation enhancement technique to the generator network to effectively capture spatial contextual information among channels. Finally, content loss is introduced to enhance the visual similarity between the input and generated images. Our method excels at removing numerous rain streaks, preserving a smooth background, and closely resembling the ground truth compared to other approaches, based on both quantitative and qualitative results, without the need for paired training images. Extensive experiments on synthetic and real-world datasets demonstrate that our approach shows superiority over most unsupervised state-of-the-art techniques, particularly on the Rain12 dataset (achieving a PSNR of 34.60 and an SSIM of 0.954) and real rainy images (achieving a PSNR of 34.17 and an SSIM of 0.953), and is highly competitive when compared to supervised methods. Moreover, the performance of our model is evaluated using RMSE, FSIM, MAE, and the correlation coefficient, achieving remarkable results that indicate a high degree of accuracy in rain removal and strong preservation of the original image’s structural details.
- 20 May 2024
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 43 (9) , 5377-5400
Design and Implementation of a Wide-Swing CMOS Multiplier for AC Source Signal Tracking and Modulation
- 20 May 2024
- Published by Springer Nature in Circuits, Systems, and Signal Processing
- Vol. 43 (9) , 5377-5400
Design and Implementation of a Wide-Swing CMOS Multiplier for AC Source Signal Tracking and Modulation