Many studies have attempted to predict chlorophyll-a concentrations using multiple regression models and validating them with a hold-out technique. In this study commonly used machine learning models, such as Support Vector Regression, Bagging, Random Forest, Extreme Gradient Boosting (XGBoost), Recurrent Neural Network (RNN), Long–Short-Term Memory (LSTM), are build new model in the Nakdong River, Korea. We employed 1–step ahead recursive prediction reflect characteristics of time series...

Abstract Diverse artificial synapse structures and materials are widely proposed for neuromorphic hardware systems beyond von Neumann architecture owing to their capability mimic complex information processing tasks such as image recognition, natural language processing, learning. Nevertheless, temporal spatial randomness in the movement of ion electron particles that exist usually prevents solid‐state‐based synaptic devices from enabling reliable modulation plasticity. An aluminum...

Network-assisted congestion control leveraging Explicit Congestion Notification (ECN) is an effective way to deal with issues on the Internet. However, we believe that existing ECN mechanism in TCP/IP protocol stack may require further optimization effectively address evolving challenges introduced by emerging technologies like immersive AR/VR applications and burgeoning field of Internet Things (IoT). To end, propose a multilevel notification called Enhanced (EECN) leverages two bits IP...

This paper aims to address sensor-related challenges in simultaneous localization and mapping (SLAM) systems, specifically within the open-source Google Cartographer project, which implements graph-based SLAM. The primary problem tackled is adaptability functionality of SLAM systems diverse robotic applications. To solve this, we developed a novel framework that integrates five additional functionalities into existing Robot Operating System (ROS). These innovations include an inertial data...

This study aims to compare the classification performance of statistical models on highly imbalanced kidney data. The health examination cohort database provided by National Health Insurance Service in Korea is utilized build with various machine learning methods. glomerular filtration rate (GFR) used diagnose chronic disease (CKD). It calculated using Modification Diet Renal Disease method and classified into five stages (1, 2, 3A 3B, 4, 5). Different CKD based estimated GFR are considered...

In edge computing, most procedures, including data collection, processing, and service provision, are handled at nodes not in the central cloud. This decreases processing burden on cloud, enabling fast responses to end-device requests addition reducing bandwidth consumption. However, have restricted storage, energy resources support computation-intensive tasks such as deep neural network (DNN) inference. this study, we analyze effect of models with single multiple local exits DNN inference...

Congestion in the network is a persistent issue that becoming more challenging with advent of technologies such as metaverse and immersive AR/VR applications. Therefore, we propose Enhanced ECN (EECN), novel network-assisted congestion feedback protocol redesigns legacy ECN. EECN notifies two levels encoded existing bits, unlike which delivers only Boolean information about congestion. Additionally, control algorithm leverages this multilevel from using EECN. Our proposed mechanism can...

As the mobile handheld devices equipped with fingerprint sensors are produced, it becomes important to protect private information of a user (i.e., image) in remote applications. In this paper, we consider possible scenarios transmit images from device server securely using both encryption and watermarking. Moreover, large-sized via restricted bandwidth, image compression should be considered. Based on experimental results, compare analyze three factors(i.e., energy consumption, watermark...

Due to the severe resource constraints of sensor hardware, energy efficiency is one critical factors for monitoring movement large-scale phenomena such as wild fire and hazardous bio-chemical material, denoted by continuous objects. In order save energy, most existing research on tracking objects focuses finding ways minimize communication cost through effective data delivery aggregation reducing number reporting nodes, not much work has been done state scheduling. Energy expected improve if...

Managing queuing delays is crucial for maintaining Quality of Service (QoS) in real-time media communications. Customizing traditional routing protocols to meet specific QoS requirements—particularly terms minimizing delay and jitter media—can be both complex time-intensive. Furthermore, these often encounter challenges when adapted vendor-specific hardware implementations. To address issues, this paper leverages the programmable features Software-Defined Networking (SDN) simplify process...

Immersive technologies such as virtual reality (VR), augmented (AR), and holograms will change users' digital experience. These immersive have a multitude of applications, including telesurgeries, teleconferencing, Internet shopping, computer games, etc. Holographic-type communication (HTC) is type media that provides an experience to users. However, HTC has different characteristics network requirements, the existing architecture transport protocols may not be able cope with stringent...

To study the scene classification in Synthetic Aperture Radar (SAR) image, a novel method based on kernel estimate, with Markov context and Dempster-Shafer evidence theory is proposed. Initially, nonparametric Probability Density Function (PDF) estimate introduced, to describe of SAR images. And then under context, both determinate PDF are adopted respectively, form primary classification. Next, results fused using an unsupervised way get Finally, regularization step used, which iterated...

For monitoring of a large-scale continuous object, large number sensor nodes might be participated with object detection and tracking. In order to reduce huge quantities data from the nodes, previous studies focus on representative selection for reporting sink. However, they simply choose representatives among without consideration node deployment environments accuracy. Hence, this paper proposes novel tracking scheme that first makes set candidates which are able show accurate boundary...

In order to monitor continuously moving phenomena such as wile fire and hazardous bio-chemical material in wireless sensor network, boundary tracking approach has been widely used by reason of its huge scale extensive diffusion property. With the scheme, energy efficiency is expected improve if only nodes near continuous object actively participate process, while other stay sleep mode for saving. this paper, we propose a predictive which uses minimum set active sensing reduce consumption....

Abstract The purpose of this study was first to extract the anthropometric data typical Korean male adults, based on three‐dimensional measured through Size Korea project. were then analyzed identify differences in characteristics between Koreans and 3D mannequinmannequins generated by digital human models. Revision equations suggested improve inaccuracy Typical adults subject body scan selected factor analysis with respect 5th, 50th, 95th percentiles. Comparisons included height length...

This research aims to find where visually impaired users appliances hard use and suggest guideline solve this issue. 181 have been surveyed, 12 selected based on disability cause classification. In a home-like environment, we had participants perform tasks which were sorted using Hierarchical task analysis six major home appliances. From found out that sometimes only provide visual information causes difficulty in sensory processing. Also, interfaces tactile/auditory feedbacks are the same...

Packet losses in the network significantly impact performance. Most TCP variants reduce transmission rate when detecting packet losses, assuming congestion, resulting lower throughput and affecting bandwidth-intensive applications like immersive applications. However, not all are due to congestion; some occur wireless link issues, which we refer as non-congestive losses. In today's hybrid Internet, packets of a single flow may traverse wired segments reach their destination. should react...

In this work, we introduce a control framework that combines model-based footstep planning with Reinforcement Learning (RL), leveraging desired patterns derived from the Linear Inverted Pendulum (LIP) dynamics. Utilizing LIP model, our method forward predicts robot states and determines foot placement given velocity commands. We then train an RL policy to track placements without following full reference motions model. This partial guidance physics model allows integrate predictive...

With the evolution of cellular networks and wireless-local-area-network-based communication technologies, services for smart device users have appeared. popularity 4G 5G, can now consume larger bandwidths than before. Consequently, demand various services, such as streaming, online games, video conferences, has increased. For improved quality experience (QoE), streaming utilize adaptive bitrate (ABR) algorithms to handle network bandwidth variations. ABR use history future prediction,...