Earticle

The urban aerial mobility (UAM) system, such as drone taxi or air taxi, is one of future on-demand transportation networks. Among them, electric vertical takeoff and landing (eVTOL) is one of UAM systems that is for identifying the locations of passengers, flying to the positions where the passengers are located, loading the passengers, and delivering the passengers to their destinations. In this paper, we propose a distributed deep reinforcement learning where the agents are formulated as eVTOL vehicles that can compute the optimal passenger transportation routes under the consideration of passenger behaviors, collisions among eVTOL, and eVTOL battery status.

Stability is the most significant component of an autonomous driving system, affecting both the lives of drivers and pedestrians and traffic flow. Reinforcement learning (RL) is a representative technology used in autonomous driving, but it has challenges because it is based on trial and error. In this letter, we propose an efficient learning approach for stable autonomous driving. The proposed deep reinforcement learning based approach can address the partially observable scenario in mixed traffic which includes both autonomous vehicles and human-driven vehicles. Simulation results show that the proposed model outperforms the control-theoretic and vanilla RL approaches. Furthermore, we confirm the effect of the sync-penalty, which teaches the agent about unsafe decisions without experiencing the accidents.

Vision, Radar, and LiDAR sensors are widely used for autonomous vehicle perception technology. Especially object detection and classification are primarily dependent on vision sensors. However, under poor lighting conditions, dazzling sunlight, or bad weather an object might be difficult to be identified with general vision sensors. In this paper, we propose a sensor fusion system that combines a thermal infrared camera and a LiDAR sensor that can reliably detect and identify objects even in environments with poor visibility, such as day or night. The proposed method obtains the external parameters of the two sensors by designing and manufacturing a 3D calibration target to externally calibrate the thermal infrared camera and the LiDAR sensor. To verify the performance, experiments were conducted in day and night environments. The proposed sensor system and fusion algorithm show that it can reliably detect and identify objects even in environments with poor visibility, such as day or night.

Within the broader context of Information and Communications Technology (ICT), the quest for reliable and scalable visual segmentation methods poses significant challenges, particularly in autonomous driving, where real-world scene complexity requires advanced solutions. To address data scarcity and improve segmentation performance, we propose a novel unsupervised domain adaptation (UDA) approach that enhances target domain learning. Our method introduces multiple perturbations consistency, leveraging spatial context within the target domain to improve recognition. By applying perturbations at input and feature levels and using a consistency loss, we enhance contextual learning. Additionally, a weight mapping technique reduces the impact of detrimental source domain information. Experimental results demonstrate that our approach outperforms baseline methods on the GTAV->Cityscapes and SYNTHIA->Cityscapes datasets.

An intelligent distributed multi-agent system (IDMS) using reinforcement learning (RL) is a challenging and intricate problem in which single or multiple agent(s) aim to achieve their specific goals (sub-goal and final goal), where they move their states in a complex and cluttered environment. The environment provided by the IDMS provides a cumulative optimal reward for each action based on the policy of the learning process. Most actions involve interacting with a given IDMS environment; therefore, it can provide the following elements: a starting agent state, multiple obstacles, agent goals, and a cluttered index. The reward in the environment is also reflected by RL-based agents, in which agents can move randomly or intelligently to reach their respective goals, to improve the agent learning performance. We extend different cases of intelligent multi-agent systems from our previous works: (a) a proposed environment-clutter-based-index for agent sub-goal selection and analysis of its effect, and (b) a newly proposed RL reward scheme based on the environmental clutter-index to identify and analyze the prerequisites and conditions for improving the overall system.

In the field of foreign language education, learner autonomy, one of the ultimate goals of education, has been greatly promoted in China ever since the late 1990s. The present study examined via a questionnaire the status quo of Chinese postgraduate students’ autonomous English learning (AEL), aiming to present a clear picture of students’ autonomous learning situation for the improving of pedagogy. 292 postgraduate students from 7 universities were asked to report their attitudes towards AEL and actual out-of-class AEL behaviors. The findings revealed that postgraduate students generally held strongly positive attitudes towards AEL but exhibited only moderate frequency of autonomous learning behaviors outside of class. Students’ attitudes and behaviors were statistically significantly correlated, but the correlation was not high enough to guarantee that positive attitudes will bring about high frequency of AEL behaviors. Except for the perception of the importance of classroom teaching and learning, students of different genders showed no significant differences in either attitudes towards AEL or AEL behaviors. Suggestions on postgraduate English teaching were subsequently discussed.

This paper reports a study that examined Chinese college English language learners’ out-of-class computer-assisted autonomous language learning (ALL) by means of questionnaires and interviews. Data were collected from 345 students enrolled in college English courses at three levels: basic, intermediate, and advanced. Their computer-assisted ALL behaviors and the value they attached to computers and networks for ALL were identified. Descriptive statistics showed that they held favorable attitudes towards ALL assisted by computers and networks, but they did not show a high frequency of ALL behaviors. The correlation analysis yielded significant positive relationships between attitudes and behaviors. Repeated one-way ANOVAs revealed significant differences in both attitudes and behaviors among the students of the three levels. The advanced students showed the most favorable attitudes and the highest frequency of computer-assisted ALL behaviors while the basic-level students demonstrated a medium degree of support for computer-assisted ALL and the least frequency of ALL behaviors. Interview results indicated that the students’ attitudes towards computer-assisted ALL and the teachers’ positive intervention in the language classes might be two important factors contributing to the differences in ALL behaviors.

In this study, we propose a deep learning-based forest road recognition technology and integrate it into an autonomous steering and speed control system for a forestry forwarder. We trained a YOLOv8 segmentation model on 640×480 pixel images collected from straight, curved, and occluded sections of forestry roads, achieving a precision of 0.997, a recall of 0.994, an F1 score of 0.995, and an average precision of 0.994. We developed a centerline extraction algorithm that calculates the lateral offset between the vehicle center and the detected road centerline using the segmented mask. This offset is then input into an electronic steering lever control system, which replaces the existing mechanical steering lever to enable rapid response. In field experiments conducted on actual forest roads, the maximum average path tracking error was 0.534 m over three trials, confirming that the vision-based approach can reliably guide the forwarder under various lighting conditions, shadows, and roadside structures. This method is expected to overcome the limitations of GNSS/INS-based autonomous driving systems, as it operates solely with cameras and single-board hardware. The research results suggest that deep learning-based image processing provides a solid foundation for fully autonomous forestry operations and has the potential to improve the productivity and safety of timber transport operations.

This study presents model-free reinforcement learn ing methods for economic and ecological adaptive cruise control (Eco-ACC) of connected and autonomous electric vehicles. For model-free optimal control of Eco-ACC, we applied two reinforcement learning methods, Deep Q-Network (DQN) and Deep Deterministic Policy Gradient (DDPG), in which deep neural networks of actors and critics were trained using IPG CarMaker simulations. For performance demonstrations, the HWFET, US06, and WLTP Class 3b driving cycles were used to simulate the front vehicle, and the energy consumptions of the host vehicle and front vehicle were compared. In high-fidelity IPG CarMaker simulations, the proposed reinforcement learning- based Eco-ACC methods demonstrated approximately 3–5% and 10–14% efficiency improvements in highway and city-highway driving scenarios, respectively, when compared with the front vehicle. A video of the CarMaker simulation is available at https://youtu.be/DIXzJxMVig8.

본 연구는 강화학습과 CycleGAN기반 데이터 증강을 결합하여 자율주행 시스템의 객체 인식 성능을 향상하는 방법을 제안한다. 비, 안개, 눈과 같은 악천후 조건의 문제를 해결하기 위해 CycleGAN을 활용하여 다양한 환경의 데이터를 생성하고, 이를 기반으로 객체 인식 모델인 YOLOv8을 학습하여 기존 데이터만 사용한 모델보다 높은 인식 정확도를 달성하였다. 또한, 강화학습을 통해 학습 과정에서 중요한 하이퍼파라미터인 학습률을 동적으로 최적화하였다. 실험 결과, 모델의 학습 효율성을 향상시켰음을 확인하였다. 본 연구 결과는 자율주행의 환경적 한계를 극복하고 더욱 신뢰할 수 있는 자율주행 시스템 연구에 기여할 것으로 기대한다.

This study introduces a method to enhance object recognition performance in autonomous driving systems by combining reinforcement learning with CycleGAN-based data augmentation. To address challenges posed by adverse weather conditions such as rain, fog, and snow, we generated diverse datasets using CycleGAN and trained YOLOv8, a state-of-the-art object detection model. This approach achieved higher recognition accuracy compared to models trained on standard datasets. Additionally, reinforcement learning was utilized to dynamically optimize critical hyperparameters, particularly the learning rate, during the training process. Experimental results demonstrate that this method improves not only the training efficiency but also the robustness of the object detection model under varying environmental conditions. The findings highlight the potential of integrating data augmentation and reinforcement learning to overcome environmental challenges in autonomous driving, contributing to the research of more reliable and adaptable autonomous systems.

Objectives: A learning organization was designed and implemented on the basis of the selection criteria and essential elements of knowledge translation theory. Methods: The learning organization was designed on the basis of biosafety harmonization criteria and risk management strategy and was implemented as the learning organization for biosafety management by the National Institute of Health, Korea Centers for Disease Control & Prevention. The effect of knowledge translation in the research institutions by evidence-based policy was verified. Results: The result of applying the knowledge translation theory involving all stakeholders showed a positive reaction in establishing and implementing biosafety management strategy and embodied risk assessment criteria and evoked sympathy with the necessity of learning and using of expert knowledge about risk assessment and risk management. All stakeholders initiated voluntarily action toward new human-network construction and communication between similar organizations. The learning organization's capability expanded the base of knowledge translation. Conclusion: These results showed that a learning organization could enhance the autonomous safety management system by diffusion of knowledge translation.

본 연구는 퍼스널 위키(Personal WIKI) 기반 학습이 학습자의 자율적 학습 태도 확립에 미치는 영 향과, 그로부터 파생되는 주체적 인격 형성의 과정을 탐색하는 데 목적이 있다. 이를 위해 Giorgi의 현상학적 연구 방법을 적용하였으며, 퍼스널 위키를 활용한 학습자의 실제 경험을 심층적으로 분석 하였다. 학습자는 위키 기반 자기 주도적 탐색 과정을 통해 외부로부터 수동적으로 전달된 지식을 수용하기보다는, 자신의 관심과 목적에 따라 학습 주제를 선정하고, 그 의미를 주관적으로 재해석하 고 전유하였다. 이러한 과정은 자율적 학습 태도의 형성을 넘어, 학습자가 자신의 사고 구조를 성찰 하고 개념을 재구성함으로써 인격적 통합과 자기 정체의 인식에 이르는 주체적 성장의 가능성을 드 러내었다. 연구 결과, 퍼스널 위키 기반 학습은 학습자가 지식의 수용자를 넘어 의미 생성자이자 삶 의 주체자로서 자리매김할 수 있는 가능성을 보여주었다. 본 연구는 퍼스널 위키가 단순한 학습 도 구를 넘어, 자율성과 주체성을 심화시키는 인격 교육적 실천으로 기능할 수 있으며, 나아가 ‘진정한 자아’와 ‘실천적 앎’을 지향하는 학습으로 전이될 수 있는 함의를 내포하고 있음을 시사한다.

This study aims to explore the impact of Personal WIKI based learning on establishing learners’ autonomous learning attitudes and the process of forming a self-directed character derived from it. To this end, Giorgi’s phenomenological research method was applied, and learners’ actual experiences using Personal WIKI were analyzed in depth. Through the WIKI-based self-directed exploration process, learners selected learning topics according to their own interests and purposes, subjectively reinterpreted and appropriated their meanings, rather than passively accepting externally delivered knowledge. This process revealed the potential for subjective growth beyond the formation of autonomous learning attitudes, enabling learners to reflect on their own thought structures, reconstruct concepts, and achieve personal integration and recognition of self identity. The research findings demonstrated that Personal WIKI based learning holds the potential to position learners not merely as knowledge, but as meaning creators and agents of their own lives. This study suggests that Personal WIKI can function not merely as learning tools but as character-educational practices that deepen subjectivity.

본고에서는 일본어수업의 학습자 내성활동을 통한 자율적 학습의 지원가능성을 밝히기 위 해 학습자의 인식과 전략사용을 중심으로 분석・고찰하였다. 우선 내성활동을 통해 얻은 학습 자의 내성내용을 분석하여 자율적 학습의 구성요인(‘동기부여・정의요인’, ‘메타인지요인’, ‘학 습전략요인’)과 하위 항목을 도출하였다. 이를 통해 내성시트에 제시된 내성항목이 암시적으 로 작용하여 학습자 스스로 위와 같은 요인을 이끌어내도록 촉진했을 가능성을 엿볼 수 있었 다. 다음으로 내성활동이 자율적 학습능력의 육성에 도움이 된다고 생각하는지, 그 이유는 무 엇인지에 대한 내성활동, 즉 자율적 학습과 연관시킬 수 있는 명시적 작용으로서의 메타내 성을 실시하였다. 그 결과 대부분의 학습자가 긍정적으로 인식하고 있었으며, 그 이유로 주로 언급되었던 메타인지가 활성화되어 자율적 학습에 대한 의식을 강화했을 가능성을 확인할 수 있었다. 끝으로 자기조정학습전략 척도에 의한 조사를 통해 내성활동에서의 의식화가 학습전 략의 사용과 같은 행동에 영향을 주는지 살펴보았다. 그 결과 내성활동 전후의 학습전략 사 용과 그 변화 양상을 알 수 있었으며 긍정적인 영향에 대한 시사를 얻을 수 있었다. 이러한 결과를 통해 이번 수업실천의 내성활동이 학습자의 자율적 학습에 대한 의식화를 촉진・강화 할 뿐만 아니라 자율적 학습전략의 사용으로 이어질 수 있다는 가능성을 실증적으로 확인할 수 있었다.

This study examines how reflection activities support autonomous learning, focusing on learners’ awareness and the use of various learning strategies. First, an examination of the reflection sheets of learners in a Japanese class revealed that the components of learner autonomy, such as “motivation or affective factors,” “metacognition factors,” and “learning strategy factors.” This indicates the possibility that reflection activities using sheets implicitly promoted the learner to elicit the above factors. Secondly, through metareflection as an explicit action that can be associated with autonomous learning, whether reflection activities helped to enhance learner autonomy, and, if so, what is the reason, was examined. The result reveals that most learners perceived positively; further, it was possible to confirm the possibility that metacognition which was specifically mentioned, was activated to reinforce the awareness of autonomous learning. Finally, to determine if consciousnessraising in reflection activities affects behavior, such as the use of learning strategies. A selfregulated learning strategy scale survey was conducted, which consisted of five factors: a) “reflecting on learning”; b) “devising learning”; c) “asking peers”; d) “developing a learning plan”; and e) “rewarding oneself.” As a result, the use of learning strategies before and after reflection activities and their changes were identified, with results that suggested a positive effect. These findings show that reflection activities may not only promote and reinforce learners’ awareness of autonomous learning but also helped to induce the use of autonomous learning strategies.

본 연구는 사이버학습에서 온라인 학습매체와 서책형 학습매체를 활용했을 때의 자율적 학습 효과성을 비교 분석하는 연구이다. 본 연구에서는 같은 학습 내용의 온․오프라인 학습 매체를 동질 집단에 제공하고, 통제된 환경에서 자율적인 학습을 시행한 후 학습 결과물인 객관적 성적 자료를 수집하여 비교 분석하는 연구 방법을 사용하였다. 전반기와 후반기 학습 결과를 취합하여 자료를 분석한 결과 두 집단 간의 유의미한 차이를 확인할 수 있었으며, 온라인 매체를 이용한 학습이 오프라인 인쇄 매체를 이용한 학습보다 자율적인 학습 면에서 효과 성이 높다고 볼 수 없다는 결론을 얻게 되었다. 이에 온라인 학습의 자율적인 학습 효과성을 높이기 위한 방안 을 제안하고자 한다.

The purpose of this paper is to compare the learning effectiveness of autonomous learning between on-line cyber learning using Internet media and off-line learning using printed media. In this study, each group is provided with the same learning contents and the same learning conditions. For autonomous learning the results revealed a significant difference between on-line cyber learning being not as effective as off-line learning. In order to improve the effectiveness of on-line cyber learning the negative aspects have been identified. These findings will help to improve the development of on-line learning effectiveness.

가상 환경 또는 실세계에서 지능적인 NPC를 위한 연구가 하드웨어나 소프트웨어 분야에서 활발히 진행되고 있다. 하드웨어 분야로는 로봇 시스템이 주를 이루고 있으며, 이는 물리적인 제약점과 노이즈의 문제 해결에 초점을 맞추고 있다는 점에서 단순 소프트웨어 분야와는 차이점이 있다. 그러나 공통적인 목표는 사람과 같은 환경에서 사람과 같이 행동하는 지능형 NPC 기술 개발이다. 본 논문은 최근 기존 연구 중에서 소프트웨어 분야와 아키텍쳐 설계를 통해 지능적인 NPC를 연구한 사례를 중점으로 조사한다. 사람과 같은 행동 결정을 유도하는 NPC에는 신경망, 유전자 알고리즘, 사례기반 추론, 학습과 같은 여러 분야에 연구가 진행되고 있다. 다양한 인공지능 알고리즘 중에서 현재 환경과 자신의 상태를 고려하여 가장 적합한 행동 결정을 유도하는 계획 기법을 기반으로 하는 연구를 조사한다. 이러한 연구는 기존의 실세계의 로봇 시스템, 가상 환경, 가상 시뮬레이션 그리고 게임의 여러 캐릭터 위주의 장르에서도 활용범위를 넓힐 수 있다.

한류열풍으로 인해 한국으로 유학 온 중국인 학생 수가 대폭 증가하였다. 그러나 이들은 대부분 20대 초반으로 자신을 스스로 관리하는 능력이 부족하다. 게다가 하루에 4시간만 수업을 받기 때문에 교사는 학습자가 여가시간에 효율적으로 한국어학습을 할 수 있도록 학습자들의 자율학습능력을 키워 주어야 한다. 학습자의 자율학습능력을 키우기 위해 고려해야 하는 요소가 많지만 그 중에서도 정의적 요인이 자율학습능력에 긍정적인 영향을 미칠 것으로 생각할 수 있다. 이에 본 연구에서는 중국인 한국어 학습자의 정의적 요인이 자율학습능력에 대해 어떤 영향을 미치는지 알아보기 위해 3가지 연구 문제로 정의적 요인과 자율학습능력과의 상관관계를 다중회귀 분석을 통해 분석하였다. 그 결과 학습태도 요인이 학습동기 요인과 불안감 요인에 비하여 자율학습에 미치는 영향이 가장 크다는 것을 알 수 있었다. 따라서 한국어교사는 수업시간에 학습자가 좋은 학습태도를 유지하도록 유도하는 것이 바람직하다고 할 수 있다.

With the impact of South Korean culture, more and more foreign students are studying abroad in South Korea. Most of these students are in the age of early twenties. Weak self-constraint ability of students in this age group, coupled with the students’ study of Korean for only four hours a day, means teachers must guide students to learn how to become more effective after school, a time that should develop the students’ self-learning ability. In order to foster students’ self-learning ability, a number of factors need to be considered. Affective factors in particular had a significant impact on the cultivation of self-learning ability. This article explores the correlation between Korean students’ affective factors and their self-learning ability. It submitted 3 research question on which the hypothesis was tested with the regression statistical methods. The results show that the learning attitude had the largest impact on developing self-learning ability than motivation and anxiety. Therefore, teachers in the classroom should encourage the students to take a correct attitude to enhance self-learning ability.

본 논문에서는 실내 환경에서 시각정보를 기반으로 출발지점에서 경유지를 거쳐 목표지점으로 최적의 경로를 찾아 자율 주행하는 바퀴달린 로봇을 구현한다. 로봇은 출발지점에서 경유지를 거쳐 목표지점으로의 최적의 경로를 딥강화학습으로 얻을 수 있다. 그러나 로봇이 구해진 경로로 자율 주행을 할 때 표면의 굴곡과 이물질 등의 외부적 요인으로 목적지까지 정확하게 주행하지 못하는 경우가 발생한다. 이에 본 연구는 카메라만 장착한 로봇이 외부 요인으로 인해 최적의 경로를 이탈할 경우 이를 인지하도록 한다. 이 인지를 토대로 로봇이 스스로 경로를 보정하고 계획된 경유지와 최종 목적지점에 도달할 수 있게 하는 알고리즘을 제안한다. 본 연구를 위해 파이캠을 탑재한 라즈베리파이와 아두이노로 제어하는 바퀴식 자율 주행 로봇이 제작되었다. 로봇은 실내환경에서 OSX 환경의 서버와 실시간 연동하면서 계획된 최적의 경로로 시험주행을 완료하였다.

In this paper, we implement a wheeled mobile robot that accurately and autonomously finds the optimal route from the starting point to the destination point based on computer vision in a complex indoor environment. We get a number of waypoints from the starting point to get the best route to the target through deep reinforcement learning. However, in the case of autonomous driving, the majority of cases do not reach their destination accurately due to external factors such as surface curvature and foreign objects. Therefore, we propose an algorithm to deepen the waypoints and destinations included in the planned route and then correct the route through the waypoint recognition while driving to reach the planned destination. We built an autonomous wheeled mobile robot controlled by Arduino and equipped with Raspberry Pi and Pycamera and tested the planned route in the indoor environment using the proposed algorithm through real-time linkage with the server in the OSX environment.

In Efforts to prevent CFIT accidents so far, have been emphasizing various education measures to minimize the occurrence of human errors, as well as enforcement measures. However, current engineering measures remain in a system (TAWS) that gives warnings before colliding with ground or obstacles, and even actual automatic avoidance maneuvers are not implemented, which has limitations that cannot prevent accidents caused by human error. Currently, various attempts are being made to apply machine learning-based artificial intelligence agent technologies to the aviation safety field. In this paper, we propose a deep reinforcement learning-based artificial intelligence agent that can recognize CFIT situations and control aircraft to avoid them in the simulation environment. It also describes the composition of the learning environment, process, and results, and finally the experimental results using the learned agent. In the future, if the results of this study are expanded to learn the horizontal and vertical terrain radar detection information and camera image information of radar in addition to the terrain database, it is expected that it will become an agent capable of performing more robust CFIT autonomous avoidance.