• Journal of Information Processing Systems
• /
• v.13 no.5
• /
• pp.1229-1242
• /
• 2017

In this paper, an interference aware distributed multi-channel MAC (IDMMAC) protocol is proposed for wireless sensor and actor networks (WSANs). The WSAN consists of a huge number of sensors and ample amount of actors. Hence, in the IDMMAC protocol a lightweight channel selection mechanism is proposed to enhance the sensor's lifetime. The IDMMAC protocol divides the beacon interval into two phases (i.e., the ad-hoc traffic indication message (ATIM) window phase and data transmission phase). When a sensor wants to transmit event information to the actor, it negotiates the maximum packet reception ratio (PRR) and the capacity channel in the ATIM window with its 1-hop sensors. The channel negotiation takes place via a control channel. To improve the packet delivery ratio of the IDMMAC protocol, each actor selects a backup cluster head (BCH) from its cluster members. The BCH is elected based on its residual energy and node degree. The BCH selection phase takes place whenever an actor wants to perform actions in the event area or it leaves the cluster to help a neighbor actor. Furthermore, an interference and throughput aware multi-channel MAC protocol is also proposed for actor-actor coordination. An actor selects a minimum interference and maximum throughput channel among the available channels to communicate with the destination actor. The performance of the proposed IDMMAC protocol is analyzed using standard network parameters, such as packet delivery ratio, end-to-end delay, and energy dissipation, in the network. The obtained simulation results indicate that the IDMMAC protocol performs well compared to the existing MAC protocols.

• The KIPS Transactions:PartC
• /
• v.17C no.2
• /
• pp.205-214
• /
• 2010

As network traffic is dramatically increasing due to the popularization of Internet, the need for application traffic classification becomes important for the effective use of network resources. In this paper, we present an application traffic classification method based on fixed IP-port information. A fixed IP-port is a {IP address, port number, transport protocol}triple dedicated to only one application, which is automatically collected from the behavior analysis of individual applications. We can classify the Internet traffic more accurately and quickly by simple packet header matching to the collected fixed IP-port information. Therefore, we can construct a lightweight, fast, and accurate real-time traffic classification system than other classification method. In this paper we propose a novel algorithm to extract the fixed IP-port information and the system architecture. Also we prove the feasibility and applicability of our proposed method by an acceptable experimental result.

• Journal of the Korea Convergence Society
• /
• v.12 no.12
• /
• pp.31-37
• /
• 2021

In this paper proposed a feature fusion and spatial attention-based modified YOLOv4S for small and occluded detection. Conventional YOLOv4S is a lightweight network and lacks feature extraction capability compared to the method of the deep network. The proposed method first combines feature maps of different scales with feature fusion to enhance semantic and low-level information. In addition expanding the receptive field with dilated convolution, the detection accuracy for small and occluded objects was improved. Second by improving the conventional spatial information with spatial attention, the detection accuracy of objects classified and occluded between objects was improved. PASCAL VOC and COCO datasets were used for quantitative evaluation of the proposed method. The proposed method improved mAP by 2.7% in the PASCAL VOC dataset and 1.8% in the COCO dataset compared to the Conventional YOLOv4S.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.4
• /
• pp.763-770
• /
• 2021

In this paper, in order to exclude the influence of BAN(Body Area Network) signals operating in the ISM band, the design and optimization process of an on-body microstrip patch antenna operating at 4.567 GHz is presented. The antenna for the monitoring of the lower legs with cancellous osteoporosis is designed to be lightweight and compact with improved return loss and bandwidth. The structure around the applied lower leg consisted of a five-layer dielectric plane. Taking into account losses, the complex dielectric constant of each layer is calculated using multi Cole-Cole model parameters, whereas a unipolar model is used for normal or osteoporotic cancellous bones. The return loss of the coaxial feed antenna on the phantom is -67.26 dB at 4.567 GHz, and in the case of osteoporosis, at the same frequency the return loss difference is 35.88 dB, and the resonance frequency difference is about 7 MHz.

• Journal of Industrial Convergence
• /
• v.20 no.6
• /
• pp.37-45
• /
• 2022

With the development of the Internet of Things (IoT), related network infrastructures require new technologies to protect against threats such as external hacking. This study proposes an L-PBFT consensus algorithm that can protect IoT networks based on a blockchain consensus algorithm. We designed a blockchain (private) model suitable for small networks, tested processing performance for ultra-small/low-power IoT devices, and verified stability. As a result of performance analysis, L-PBFT proved that at least the number of nodes complies with the operation of the consensus algorithm(minimum 14%, maximum 29%) and establishes a trust network(separation of secure channels) different from existing security protocols. This study is a 4th industry convergence research and will be a foundation technology that will help develop IoT device security products in the future.

• Journal of Internet of Things and Convergence
• /
• v.8 no.2
• /
• pp.55-59
• /
• 2022

Recently, IoT technology has great attention and plays a key role in 4^th industrial revolution in order to design customized products and services. Additive Manufacturing (AM) is applied to fabricate IoT sensor directly or IoT sensor embedded structure. Also, design methods for AM are developing to consolidate various parts of IoT devices. Part consolidation leads to assembly time and cost reduction, reliability improvement, and lightweight. Therefore, a design method was proposed to guide designers to consolidate parts. The design method helps designers to define product architecture that consists of functions and function-part relations. The product architecture is converted to a network graph and then Girvan Newman algorithm is applied to cluster the graph network. Parts in clusters are candidates for part consolidation. To demonstrate the usefulness of the proposed design method, a case study was performed with e-bike fabricated by additive manufacturing.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.9
• /
• pp.1330-1339
• /
• 2022

3D object detection generally aims to detect relatively large data such as automobiles, buses, persons, furniture, etc, so it is vulnerable to small object detection. In addition, in an environment with limited resources such as embedded devices, it is difficult to apply the model because of the huge amount of computation. In this paper, the accuracy of small object detection was improved by focusing on local features using only one layer, and the inference speed was improved through the proposed knowledge distillation method from large pre-trained network to small network and adaptive quantization method according to the parameter size. The proposed model was evaluated using SUN RGB-D Val and self-made apple tree data set. Finally, it achieved the accuracy performance of 62.04% at mAP@0.25 and 47.1% at mAP@0.5, and the inference speed was 120.5 scenes per sec, showing a fast real-time processing speed.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.3
• /
• pp.9-14
• /
• 2022

In the IoT(internet of things) where various devices can be connected, failure of essential devices may lead to a lot of economic and life losses. For reducing the losses, fault diagnosis techniques have been considered an essential part of IoT. In this paper, the method based on a graph neural network is proposed for determining fault and classifying types by extracting features from vibration data of systems. For training of the deep learning model, fault dataset are used as input data obtained from the CWRU(case western reserve university). To validate the classification performance of the proposed model, a conventional CNN(convolutional neural networks)-based fault classification model is compared with the proposed model. From the simulation results, it was confirmed that the classification performance of the proposed model outweighed the conventional model by up to 5% in the unevenly distributed data. The classification runtime can be improved by lightweight the proposed model in future works.

• International Journal of Computer Science & Network Security
• /
• v.22 no.12
• /
• pp.37-50
• /
• 2022

Modern supply chains include multiple activities from collecting raw materials to transferring final products. These activities involve many parties who share a huge amount of valuable data, which makes managing supply chain systems a challenging task. Current supply chain management (SCM) systems adopt digital technologies such as the Internet of Things (IoT) and blockchain for optimization purposes. Although these technologies can significantly enhance SCM systems, they have their own limitations that directly affect SCM systems. Security, performance, and scalability are essential components of SCM systems. Yet, confidentiality and scalability are one of blockchain's main limitations. Moreover, IoT devices are lightweight and have limited power and storage. These limitations should be considered when developing blockchain-based IoT-SCM systems. In this paper, the requirements of efficient supply chain systems are analyzed and the role of both IoT and blockchain technologies in providing each requirement are discussed. The limitations of blockchain and the challenges of IoT integration are investigated. The limitations of current literature in the same field are identified, and a secure and scalable blockchain-based IoT-SCM system is proposed. The proposed solution employs a Hyperledger fabric blockchain platform and tackles confidentiality by implementing private data collection to achieve confidentiality without decreasing performance. Moreover, the proposed framework integrates IoT data to stream live data without consuming its limited resources and implements a dualstorge model to support supply chain scalability. The proposed framework is evaluated in terms of security, throughput, and latency. The results demonstrate that the proposed framework maintains confidentiality, integrity, and availability of on-chain and off-chain supply chain data. It achieved better performance through 31.2% and 18% increases in read operation throughput and write operation throughput, respectively. Furthermore, it decreased the write operation latency by 83.3%.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.3
• /
• pp.89-100
• /
• 2023

Recently, on-device artificial intelligence (AI) solutions using mobile devices and embedded edge devices have emerged in various fields, such as computer vision, to address network traffic burdens, low-energy operations, and security problems. Although vision transformer deep learning models have outperformed conventional convolutional neural network (CNN) models in computer vision, they require more computations and parameters than CNN models. Thus, they are not directly applicable to embedded edge devices with limited hardware resources. Many researchers have proposed various model compression methods or lightweight architectures for vision transformers; however, there are only a few studies evaluating the effects of model compression techniques of vision transformers on performance. Regarding this problem, this paper presents a performance evaluation of vision transformers on embedded platforms. We investigated the behaviors of three vision transformers: DeiT, LeViT, and MobileViT. Each model performance was evaluated by accuracy and inference time on edge devices using the ImageNet dataset. We assessed the effects of the quantization method applied to the models on latency enhancement and accuracy degradation by profiling the proportion of response time occupied by major operations. In addition, we evaluated the performance of each model on GPU and EdgeTPU-based edge devices. In our experimental results, LeViT showed the best performance in CPU-based edge devices, and DeiT-small showed the highest performance improvement in GPU-based edge devices. In addition, only MobileViT models showed performance improvement on EdgeTPU. Summarizing the analysis results through profiling, the degree of performance improvement of each vision transformer model was highly dependent on the proportion of parts that could be optimized in the target edge device. In summary, to apply vision transformers to on-device AI solutions, either proper operation composition and optimizations specific to target edge devices must be considered.