Article
Computer Science, Information Systems
Seyed Yahya Nikouei, Yu Chen, Sejun Song, Baek-Young Choi, Timothy R. Faughnan
Summary: This paper proposes an iSENSE system to explore the feasibility of moving ML to the edge, and introduces lightweight models and a hybrid lightweight tracking algorithm to improve performance. Experimental results demonstrate that the system can track human objects in real-time on edge devices with limited resources.
IEEE TRANSACTIONS ON SERVICES COMPUTING
(2021)
Article
Engineering, Electrical & Electronic
Ao Ding, Yong Qin, Biao Wang, Limin Jia, Xiaoqing Cheng
Summary: Intelligent fault diagnosis of train bogie bearings based on edge computing is a promising technology to ensure the safety and reliability of train operation, which can give fault diagnosis systems better real-time performance and lower communication costs. This article proposes a new multiscale lightweight network with adaptive pruning for the intelligent diagnosis fault of train bogie bearings in edge computing scenarios. Experimental results demonstrate that the accuracy and complexity of the proposed network are superior to other state-of-the-art lightweight bearing fault diagnosis networks under varying operating conditions.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Agriculture, Multidisciplinary
Shilin Li, Shujuan Zhang, Jianxin Xue, Haixia Sun
Summary: The efficient detection of the flat jujube in a complex natural environment is significant for intelligent agricultural operations. This study proposes an improved lightweight algorithm based on You Only Look Once (YOLOv5) to address the issues of low detection efficiency and complex algorithm deployment. Experimental results show that the improved model achieves higher accuracy and reduces computational cost compared to mainstream target detection algorithms.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Mathematics
Mohammad Hijji, Hikmat Yar, Fath U. Min Ullah, Mohammed M. Alwakeel, Rafika Harrabi, Fahad Aradah, Faouzi Alaya Cheikh, Khan Muhammad, Muhammad Sajjad
Summary: Nowadays, people prefer to use private transport due to its low cost, comfortable ride, and personal preferences, resulting in a reduction in the use of public transportation. However, the use of personal transport has led to numerous road accidents due to drivers' conditions such as drowsiness, stress, tiredness, and age. To address this issue, an efficient deep learning-assisted intelligent fatigue and age detection system (FADS) was proposed to detect and identify different states of the driver. The system outperformed state-of-the-art techniques in experiments conducted on custom and publicly available datasets.
Article
Computer Science, Information Systems
Zongpu Zhang, Tao Song, Liwei Lin, Yang Hua, Xufeng He, Zhengui Xue, Ruhui Ma, Haibing Guan
Summary: The paper proposes a heterogeneous distributed deep neural network (HDDNN) framework for ubiquitous intelligent computing. It optimizes the utilization of hierarchical distributed systems for DNN and tailors DNN for real-world distributed systems, resulting in low response time, high performance, and better user experience.
IEEE TRANSACTIONS ON BIG DATA
(2022)
Article
Chemistry, Analytical
Li Cheng, Xuemin Zheng, Mingxin Zhao, Runjiang Dou, Shuangming Yu, Nanjian Wu, Liyuan Liu
Summary: This paper introduces a lightweight and hardware-friendly visual object-tracking network called SiamMixer. It reduces memory usage and improves accuracy through patch-by-patch inference and feature map merging encoding.
Article
Engineering, Electrical & Electronic
Yuefan Zhu, Xiaoying Liu
Summary: Wind turbine blade condition monitoring is crucial for preventing downtime and loss due to harsh environments and weather conditions. A lightweight convolutional neural network called WTBMobileNet is proposed for wind turbine blades, based on spectrograms, to reduce computation and size while maintaining high accuracy. Data augmentation is analyzed, and WTBMobileNet is shown to have interpretability and transparency for reliable deployment. WTBMobileNet is also explored in drone image classification and spectrogram object detection, demonstrating its potential applications beyond spectrogram classification.
Article
Automation & Control Systems
Zhengqiao Luo, Chuan Lin, Fuzhang Li, Yongcai Pan
Summary: Edge detection is fundamental to advanced computer vision tasks, and deep learning-based methods are computationally expensive. Researchers have developed lightweight CNNs to address this issue by simulating the interaction between two visual pathways in the biological visual system. The proposed bio-inspired lightweight edge detection CNN achieves competitive results with lower computational costs.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Agronomy
Shuzhi Su, Runbin Chen, Xianjin Fang, Yanmin Zhu, Tian Zhang, Zengbao Xu
Summary: This study proposes a novel lightweight grape detection method, which utilizes Uniformer as the backbone network and fuses low-resolution feature maps and high-resolution feature maps through BiPANet, achieving improved detection performance. Experimental results show that the method performs better in terms of precision and recall in grape detection tasks.
Article
Computer Science, Information Systems
Longyu Zhou, Supeng Leng, Qiang Liu, Haoye Chai, Jihua Zhou
Summary: This article proposes a new hierarchical tracking structure based on edge intelligence technology, which integrates the computing resource of mobile nodes and edge servers for real-time tracking. Furthermore, it introduces a long-term dynamic resource allocation algorithm to obtain an optimal resource scheduling solution for accurate and consecutive tracking.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Artificial Intelligence
D. Santhadevi, B. Janet
Summary: An intelligent threat detection framework for IoT networks is proposed, considering four key layout ideas. The Hybrid Stacked Deep Learning (HSDL) model is introduced based on these concepts, and is tested using three benchmark datasets.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Linrun Qiu, Dongbo Zhang, Yuan Tian, Najla Al-Nabhan
Summary: This study develops a target detection algorithm based on deep learning technologies, which improves recognition precision and model's convergence speed through fused edge features. The experimental results demonstrate efficient recognition rate and real-time performance of the algorithm.
JOURNAL OF SUPERCOMPUTING
(2021)
Article
Computer Science, Artificial Intelligence
Jielei Wang, Zongyong Cui, Ting Jiang, Changjie Cao, Zongjie Cao
Summary: This paper proposes a lightweight network model for ship target detection in synthetic aperture radar (SAR) imagery. The network structure optimization algorithm based on the multi-objective firefly algorithm (NOFA) is designed to encode the filters of a well-performing ship target detection network into a list of probabilities. The multi-objective firefly optimization algorithm (MFA) further optimizes the probability list to output a set of lightweight network encodings.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Computer Science, Hardware & Architecture
D. Santhadevi, B. Janet
Summary: Cyber-attacks on IoT devices are increasing due to the lack of security measures and the growing number of connected devices. The Edge Service plays a crucial role in detecting and preventing malware attacks by receiving and processing network traffic data, developing deep learning networks for malware detection, and analyzing real-time network traffic to identify patterns and anomalies.
JOURNAL OF SUPERCOMPUTING
(2023)
Article
Computer Science, Hardware & Architecture
Kutub Thakur, Hamed Alqahtani, Gulshan Kumar
Summary: The intelligent system IDGADS is capable of quickly detecting algorithmically generated domains with 99% accuracy based on easy-to-compute features of real domain name system (DNS) traffic. It can serve as the first line of defense in a security stack for validating DNS queries.
COMPUTERS & ELECTRICAL ENGINEERING
(2021)
Article
Engineering, Electrical & Electronic
Tiantian Tang, Donglai Jiao, Tao Chen, Guan Gui
Summary: This article proposes a data augmentation algorithm based on the K-means clustering algorithm and synthetic minority oversampling technique (SMOTE) to enhance medium- and long-term precipitation forecasting. By constructing multiple models and comparing the results, it is found that the SMOTE-km-XGB method is more suitable for precipitation forecasting.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Haitao Zhao, Jiawen Tang, Bamidele Adebisi, Tomoaki Ohtsuki, Guan Gui, Hongbo Zhu
Summary: An adaptive vehicle clustering algorithm based on fuzzy C-means algorithm is proposed in this paper, aiming to minimize the power consumption of vehicles. By dynamically allocating computing resources and selecting clustering heads, the algorithm can decrease power consumption while satisfying vehicle delay requirements, as confirmed by simulation results.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2022)
Article
Computer Science, Hardware & Architecture
Xingzhi Chang, Wei Liu, Chuan Zhu, Xiaohua Zou, Guan Gui
Summary: In this paper, a bilayer Markov random field (BMRF) method is proposed to address the issue of false detections caused by the lack of edge information in patterned fabric defect detection. The proposed method utilizes a constraint layer and a data layer, along with a new potential function and parameter estimation method, to achieve high recall rates on star-, box-, and dot-patterned fabrics.
JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS
(2022)
Review
Chemistry, Analytical
Anand Kumar, Sudhan Majhi, Guan Gui, Hsiao-Chun Wu, Chau Yuen
Summary: This paper discusses the importance and applications of blind modulation classification in future wireless communications, and provides a comprehensive overview of various statistical and machine learning techniques. The advantages and limitations of these methods are emphasized, and comparisons are made through simulations and experiments. Future research directions in blind MC are also discussed.
Article
Computer Science, Information Systems
Pengyu Wang, Yufan Cheng, Binhong Dong, Guan Gui
Summary: This letter proposes optimized binarized neural networks (BNNs) for wireless interference identification (WII) by constraining weights and activations to binary values, achieving extreme quantization. A novel approximation method is introduced to overcome the difficulty in propagating gradients during back-propagation. Additionally, two techniques are proposed to minimize quantization noise and create multiple routes for parameter updates, resulting in improved performance.
IEEE WIRELESS COMMUNICATIONS LETTERS
(2022)
Article
Engineering, Civil
Yuchao Chen, Jinlong Sun, Yun Lin, Guan Gui, Hikmet Sari
Summary: This paper presents a novel aircraft coordinate prediction hybrid model based on deep learning, which combines inception modules and LSTM modules to extract spatial and temporal features of dataset. The model uses ADS-B signal strength instead of specific information to obtain aircraft coordinates, sacrificing precision for reliability. Experimental results show that the proposed 2-Inception-LSTM model is optimal for positioning reliability, suitable for scenarios where high accuracy of aircraft coordinates is not required.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Environmental Sciences
Wenmei Li, Huaihuai Chen, Qing Liu, Haiyan Liu, Yu Wang, Guan Gui
Summary: This article introduces a solution to the classification of hyperspectral remote sensing images by introducing an attention mechanism and depthwise separable convolution to a three-dimensional convolutional neural network. The proposed models, 3DCNN-AM and 3DCNN-AM-DSC, have been shown to improve classification accuracy and reduce computing time.
Article
Engineering, Electrical & Electronic
Yiyang Ni, Xiaoqing Li, Haitao Zhao, Jie Yang, Wenchao Xia, Guan Gui
Summary: The study introduces an effective hybrid V2V/V2I transmission method based on a neural network to minimize transmission latency by predicting vehicle arrival rate and constructing an objective function, leading to significantly lower overall transmission latency compared to pure V2I transmission methods.
PHYSICAL COMMUNICATION
(2022)
Article
Computer Science, Information Systems
Segun Popoola, Ruth Ande, Bamidele Adebisi, Guan Gui, Mohammad Hammoudeh, Olamide Jogunola
Summary: This article proposes a federated deep learning method for zero-day botnet attack detection in IoT-edge devices. The method generates a global model by coordinating the training of independent models in multiple IoT-edge devices, achieving high-performance classification detection of zero-day botnet attacks and ensuring data privacy and security.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Cheng Cheng, Liang Guo, Tong Wu, Jinlong Sun, Guan Gui, Bamidele Adebisi, Haris Gacanin, Hikmet Sari
Summary: This article introduces a conflict detection algorithm based on ADS-B technology for aerial vehicles and further improves flight safety and conflict detection by predicting the trajectories of aerial vehicles.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Ruijie Zhao, Guan Gui, Zhi Xue, Jie Yin, Tomoaki Ohtsuki, Bamidele Adebisi, Haris Gacanin
Summary: This article proposes a lightweight deep neural network (LNN) based NID method for IoT, which achieves excellent classification performance and is suitable for classifying IoT traffic through feature dimensionality reduction and low computational cost feature extraction.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Telecommunications
Guanghui Fan, Jinlong Sun, Guan Gui, Haris Gacanin, Bamidele Adebisi, Tomoaki Ohtsuki
Summary: Due to the lack of channel reciprocity in FDD massive MIMO systems, downlink CSI needs to be continuously fed back to the base station from the user equipment, consuming bandwidth resources. This paper proposes a fully convolutional neural network for compressing and decompressing the downlink CSI. Experimental results demonstrate that the proposed method outperforms the baseline in terms of reconstruction performance and reduction of storage and computational overhead, and is robust to quantization error in real feedback scenarios.
IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND NETWORKING
(2022)
Article
Optics
Ehsan Mostafapour, Changiz Ghobadi, Javad Nourinia, Seyed Sadra Kashef, Guan Gui
Summary: Mobile diffusion adaptive networks are distributed sensor networks used for tracking and localization. Link noise during data exchange between the nodes can have a significant impact on the network performance, causing large convergence errors. This paper models and analyzes the performance of mobile adaptive networks using multiplicative noise in visible light communication (VLC).
OPTICS COMMUNICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Tiantian Tang, Tao Chen, Guan Gui
Summary: Satellite precipitation products (SPPs) are significant data sources in hydrometeorology, especially for ungauged or sparsely gauged basins. However, these products have varying degrees of uncertainty and their applicability may differ in different regions. This study performs statistical evaluations and improves the accuracy of five SPPs using a merging model. The utility of the precipitation sets is investigated using a hydrological model, and the results show significant improvements in two basins.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Ziteng Wang, Wenmei Li, Guan Gui
Summary: Explored an image classification method for high spatial resolution remote sensing images. Utilized transfer learning to improve performance with a small sample size. Experimental results demonstrated that increasing the sample size can stabilize the classification performance of the model.
ADVANCED HYBRID INFORMATION PROCESSING, PT I
(2022)