ETRI Journal最新文献

Thrombocytopenia is a common complication among critically ill patients. To enable early prediction, we conducted a retrospective study using five machine learning (ML) models developed with a sequence embedding approach that integrates temporal medication and diagnostic data. Models were trained on the MIMIC-IV database and evaluated on the eICU database. We propose a novel sequence feature fusion method combining explicit and implicit features with embeddings for ICD codes and drug sequences to capture complex interactions. To our knowledge, this is the first study to make continuous predictions for ICU patients until thrombocytopenia onset. Model performance was assessed using AUC; t-SNE and SHAP were used to evaluate feature importance. XGBoost with sequence feature fusion performed best, achieving AUCs of 0.80, 0.85, and 0.92 at ICU admission, and 72 h and 24 h before onset, respectively. Platelet count, phosphate, and lactate were the top predictors. These findings demonstrate that ML models with sequence embeddings can effectively predict thrombocytopenia by capturing temporal patterns in patient data.

Texts are widely used in natural language processing. However, such applications are vulnerable to adversarial attacks. Existing research attempts to artificially add semantically meaningless word-, character-, or sentence-level perturbations, which compromise the syntax and consistency of texts. However, they fail to ensure high-quality outputs. Therefore, we propose an attack model for generating adversarial samples using policy gradients and a generative adversarial network. In our model, first, a Seq2Seq encoder is used to generate sentences, mapping discrete text data into continuous hidden space vectors and then transforming them into adversarial text samples. Second, to emphasize semantics, we compute the cosine similarity or BERT-based semantic similarity between the original and adversarial texts for reward calculation. Finally, a policy gradient is applied to optimize the parameters. Experiments show that, while maintaining a semantic similarity above 0.8, our BERT-based method reduces classification accuracy by 51.77% on the DBpedia dataset. Our cosine similarity-based method requires only one-third to one-half the runtime of the baseline approach.

Three-dimensional (3D) geospatial technologies are essential in urban digital twins, smart cities, and metaverse. Rendering large-scale terrain data, often exceeding tens of terabytes, presents challenges. While planetary-scale platforms, like Google Earth and Cesium stream data, the streaming of data and the use of regular grid-type digital elevation models lead to cracks among tiles with different levels of detail. This paper proposes a novel dynamic tile-map generation method to eliminate these cracks. Unlike existing methods, our approach leverages tile subindex information to efficiently construct a tile adjacency map, significant reducing the search space for neighboring tiles and eliminating the need for prior knowledge of the terrain tile structure. Furthermore, our approach is robust to data loss, mitigating cracks caused by missing or incomplete tiles. Compared with existing root-down search methods, our method reduces processing time by 1–5 ms per frame and decreases the number of tile-to-tile links by a factor of 3–5, as demonstrated by experimental results.

Coded caching reduces the communication load substantially, exploiting the caches of end devices to generate multicast opportunities during the transmission phase. To address user-request privacy, we propose a decentralized coding caching method that focuses on protecting user privacy. This method involves creating file subpackages for users to cache linear combinations of files. We also expand the key scheme for decentralized situations, ensuring that files shared among users do not exceed each user's cache size. We make sure that the unencoded part of each packet in the user cache is larger than the size of the cached file after being cut, determining the range of values for the file allocation coefficient, θ. With fixed N and M, we can calculate that the load is a convex function of θ. Through mathematical analysis, we can determine the worst case load scenario. Subsequent simulation results unequivocally demonstrate the capability of the proposed scheme to fulfill any file request from users, all while achieving a communication load comparable to that of an enhanced distributed nonprivate cache scheme.

In today's technological landscape, the rapid and widespread adoption of new technologies is crucial to enhance the capabilities, robustness, and efficiency of military defense and disaster response operations. Technologies such as artificial intelligence, mobile communication, and the Internet of Things have enriched battlefield communication, surveillance, tactical decision-making, and early warning systems. This trend is common across various fields, including disaster response technologies, and has led to considerable improvements in disaster prediction, mitigation, response, and recovery applications.

The emergence of new technologies has resulted in dramatic changes in the operational environment. For example, the increasing diversity of connections between combat/rescue equipment, weaponry, and operational headquarters imposes complex communication requirements related to availability, reliability, and latency, as well as the need for safe processing of unprecedented volumes of data. Conversely, responses to disasters must consider their potential impacts, including their high frequency, widespread damage, and global scale. Additionally, preemptive interventions that allow for accurate forecasting of disasters are essential for modern disaster response. Overall, myriad factors collectively contribute to the complexity of developing efficient solutions for military defense and disaster response applications.

The Electronics and Telecommunications Research Institute (ETRI) Journal is a peer-reviewed open-access journal launched in 1993 and published bimonthly by ETRI (Republic of Korea), aiming to promote worldwide academic exchange in the fields of information, telecommunications, and electronics. This special issue explores recent research trends in the technological advances driving the digital transformation of military defense and disaster response systems. It presents notable, cutting-edge studies aimed at improving the efficiency, safety, and real-time responsiveness of these critical domains. Given the central role of technologies such as virtual training, robotic navigation, drone countermeasures, and secure communications in the modernization of defense operations, the contributions in this special issue offer valuable insights into the future direction of digitalized military defense and disaster response strategies. Accordingly, we have selected eight critical papers on three aspects of military defense and disaster response technology for this special issue. A brief review regarding commitments for this special issue follows.

The first invited paper [1], entitled “Next-generation wireless communication technologies for improved disaster response and management” by Song et al., introduces next-generation wireless communication technologies that can improve disaster response and management. This study proposes an integrated disaster-response communication framework with the potential to achieve ul

The National Fire Agency (NFA) and National Police Agency (NPA) have defined risk levels based on the severity of disasters. Risk-level data possess the characteristics of ordinal data such as NPA's Emergency Service Response Code (ESRC) data, which are classified based on their magnitudes (from C0 to C4). In this study, we propose a distance mean-square (DiMS) loss function to improve the accuracy of ordinal data classification. The DiMS loss function calculates loss values based on the distances between the predicted and true labels: value distances (commonly used in regression analysis for magnitude data) and probability distances (typically used in classification analysis). Therefore, the DiMS loss function contributes to improved accuracy when classifying ordinal data, such as ESRC. In addition, using the DiMS loss function, we achieved state-of-the-art performance in classifying the SST-5 data, which is a representative ordinal dataset. The DiMS loss function for ordinal classification enabled accurate risk recognition. Thus, accurate risk recognition using the DiMS loss function enhances disaster response.

Recently, numerous studies have been conducted to incorporate the knowledge of massive text corpora into speech recognition via text to speech (TTS). However, the distribution mismatch between synthetic and real speech has always been an issue. In this paper, we analyzed how these mismatches affect the acoustic and linguistic aspects of automatic speech recognition (ASR) performance. For acoustics, we divided the acoustic mismatch into TTS-related and non-TTS-related and analyzed how each acoustic mismatch affected ASR performance. Next, from a linguistic perspective, we experimented to determine how synthetic speech from a large text corpus affects the performance of speech recognition in various domains. The experimental results show that (i) substitution errors, which are the bulk of the recognition errors in ASR trained on synthetic speech data, are affected by the prosody mismatch between synthetic and real speech; (ii) pretraining ASR with synthetic speech data first and performing transfer learning with real speech outperformed training in the reverse order; and (iii) pretraining with a large amount of synthetic speech improves performance further in language model shallow fusion.

Edge computing deploys computing resources to the network edge to diminish task processing delays and power consumption. However, in traffic video surveillance systems, vehicular movement can lead to service interruptions. Moreover, the low credibility of the edge systems affects the success rate of edge offloading. To address these issues, we propose a secure task offloading scheme for traffic video surveillance. This scheme comprehensively evaluates trust values by integrating direct and indirect trust values, selecting nodes based on predefined trust thresholds, and achieving secure offloading and migration with short delays and energy consumption. To model the service migration problem, we adopted a Markov decision process-based approach and employed a $�Q$-learning algorithm to determine the optimal solution, thereby establishing an effective migration path. Simulation results demonstrate that the proposed scheme yields higher comprehensive trust values and improved reliability. Furthermore, while ensuring seamless task migration, the migration success rate exceeded 90%, with the energy consumption reduced by more than 9.8%.

The complexity of modern battlefields demands advanced military training systems that prepare forces for realistic scenarios. Even though traditional training methods are effective, they are costly and time consuming and are associated with safety risks. Virtual training systems offer a safer and more cost-effective alternative; however, current solutions often compromise realism because of the need for multiple sensors and wearable devices that can diminish immersion. To address these limitations, we propose a marker-based, adaptive, virtual military training system (MAVMTS) that enhances realism by using a minimal set of fiducial markers and multiview cameras to estimate the trainee's posture and weapon orientation without cumbersome wearables. The system integrates action recognition to generate responsive virtual adversaries, thereby creating dynamic and immersive training environments. MAVMTS reduces considerably the equipment burden and enhances the realism of virtual military training, thereby offering a more effective solution for preparing personnel for modern warfare.

This paper introduces next-generation wireless communication technologies applicable to disaster response. With recent technological advances—such as 5G new radio vehicle-to-everything (NR-V2X)—real-time communication, decision-making, and emergency operations can be enhanced through ultralow latency and high-speed transmission. Additionally, several technologies—such as autonomous mobility, drones, and Internet of Things (IoT)-based sensor networks—can improve rescue operations and mitigate human risks during disasters. Moreover, future communication technologies—including platooning, digital twins, and cell on wheels (COW)—can support situational awareness and rapid decision-making in disaster environments. Consequently, we propose industrial applications and standardization strategies to effectively utilize next-generation communication technologies for disaster response. Lastly, we analyze current wireless technologies and suggest future research directions for optimizing disaster preparedness and response.