Earticle

In this paper, we propose a deep reinforcement learning (DRL) based unmanned aerial vehicles (UAV)-assisted trajectory optimization for ultra-dense small cell networks. We assume that each UAV is equipped with a sensing radio to obtain distance information to the UEs and other UAVs in the network which are used to update the UAV’s trajectory. The proposed DRL-based system selects the optimal joint control actions for the UAVs that maximizes the system sum-rate. The simulation results show that the proposed DRL-based UAV controller provides fast UAV placement in the network with a high system performance when compared with the benchmark schemes.

In this paper, we will consider the problem of multi-unmanned aerial vehicles (UAVs) deployment among users’ hotspots. UAVs carry reconfigurable intelligent surface (RIS) boards to strengthen millimeter wave (mmWave) coverage at these hotspots. UAVs should maximize hotspots’ sum data rates while minimizing their flights’ cost, i.e., hovering and flying energy consumptions. Moreover, collisions should be avoided among UAVs, i.e., one hotspot should be served by only one UAV at a time. In this paper, this problem is considered as a multi-player multi-armed bandit (MP-MAB) game with budget constraint and collision avoidance. In this context, budget constraint Thompson sampling (TS) with collision avoidance MAB algorithm (BTSCA-MAB) is proposed to efficiently implement the formulated MP-MAB game. Numerical analysis confirms the superior performance of the proposed BTSCA-MAB over other benchmarks.

This paper proposes a method for characterizing user mobility levels in multiple unmanned aerial vehicles (UAV)-assisted networks. In a dynamic environment with varying degrees of mobile ground users, optimizing the placement of UAVs is important in improving network throughput. Moreover, for integrated sensing and communication (ISAC) enabled UAVs, the resource allocation for sensing and communication relies on the dynamic nature of the network environment. To keep track of the changes in the environment UAVs are required to continuously update their trajectory, hence, characterizing the users’ mobility level is an important tool to improve the UAV trajectory optimization. In this paper, a mobility-aware resource allocation for joint sensing and communication is proposed. Our results demonstrate that the proposed algorithm can improve the resource allocation between sensing and communication in an ISAC-enabled UAV-assisted network.

This paper presents a real-time vehicle detection and tracking system using an unmanned aerial vehicle (UAV) to address challenges in dynamic urban environments. The system combines a convolutional neural network (CNN) for vehicle detection with a deep Q-network (DQN)-based navigation policy for continuous tracking. Input images are enhanced using contrast limited adaptive histogram equalization (CLAHE) and unsharp masking. The CNN jointly predicts vehicle center coordinates and probabilistic heatmaps, while a self-attention module captures long-range spatial dependencies to improve detection under clutter and occlusion. The DQN is trained on multi-step spatiotemporal states to learn optimal UAV movement strategies under diverse weather and structural conditions. Experiments conducted in a three-dimensional (3D) urban simulation environment using Unity’s machine learning agents (ML-Agents) show that the self-attention design reduced pixel-level localization error by about 7%, and the DQN-based tracking policy achieved stable convergence after approximately 2000?3000 episodes. These results demonstrate high tracking accuracy and system stability, highlighting the potential of the proposed approach for real-world UAV-based traffic monitoring applications.

무인기 기술의 발전에 따라 군사뿐만 아니라 다양한 민간 분야에서도 무인기 탐지 및 위치 추정 기술의 중요성이 부 각되고 있다. 딥러닝 기술의 발전으로 무인기 탐지 성능은 크게 향상되었으나, 정밀한 위치 추정에 대한 연구는 상 대적으로 미진하다. 본 연구에서는 RGB와 LiDAR로 구성된 다중 모달리티 센서팩을 설계하고, 이를 기반으로 한 데이터 세트를 수집하여 무인기 위치 추정을 위한 두 가지 시스템을 제안한다. 첫 번째 시스템은 LiDAR 포인트 클 라우드를 RGB 이미지에서의 객체 탐지 결과에 투영하여 위치를 추정하며, 두 번째는 RGB와 LiDAR 데이터를 결 합해 3D 객체를 직접 탐지한 후 위치를 산출한다. 연산 복잡도가 다소 증가하나, 두 번째 시스템은 UTM 좌표계 기준으로 동방(x) 0.77m, 북방(y) 2.07m, 평균 2.21m의 RMSE를 기록하며 더욱 우수한 정확도를 보였다. 본 연구는 다중 모달리티 기반의 무인기 위치 추정 기술이 실현 가능하고 효과적임을 보여주며, 향후 군사 및 민간 분 야에서의 실질적 응용 가능성을 제시한다.

As UAV technology rapidly evolves, the need for accurate detection and localization has become critical — not only for military and defense but also for a wide range of civilian applications. While recent advances in deep learning have significantly improved UAV detection, precise localization remains a less explored area. In this study, we design a multi-modal sensor pack combining RGB and LiDAR and build a custom dataset to develop and evaluate two UAV localization systems. The first system projects 3D LiDAR point clouds onto 2D object detections from RGB images. The second system fuses RGB and LiDAR inputs for 3D object detection, directly estimating UAV positions in 3D space. Although the second approach is more computationally intensive, it achieves better accuracy, with RMSE values of 0.77 m (x-axis), 2.07 m (y-axis), and 2.21 m overall in UTM coordinates. These findings confirm the promise of multi-modal UAV localization for both military and civilian applications.

4차 산업혁명 시대의 미래 전장은 초연결, 고속기동화된 무기체계로 다영역작전을 수행할 것이다. 이러한 미래전 양상의 변화에 대비하기 위해 우리군은 다양한 유·무인 무기체계를 개발하고 이들의 기동간 통신 지원이 가능한 다계층 전술네트워크 구성을 위하여 노력하고 있다. 그러나 현재의 전술네트워크는 단일 계층에서 단일빔 안테나를 활용한 1:1 고속링크 또는 무지향성 안테나를 활용한 1:N 저속링크를 운용하고 있어 기동간 통신 지원이 제한된다. 즉 미래전 대비 를 위한 다계층 전술네트워크를 효율적으로 구성하기 위해서 다중빔 안테나의 운용이 필요하다. 특히 공중계층의 공중 중계 무인기는 비행체 특징에 따라 다중빔 안테나의 운용 방법이 달라진다. 따라서 본 논문에서는 다계층 전술네트워크 의 효율적인 운용을 위하여 공중계층에 필요한 다중빔 안테나 운용 시나리오와 고려 요소, 회전익과 고정익 비행체의 특징을 살펴보고 이를 토대로 다중빔 안테나의 공중중계 무인기 설치 위치 및 운용 방안을 회전익과 고정익 비행체로 구분하여 제시한다.

In the era of the Fourth Industrial Revolution, the future battlefield will carry out multi-area operations with hyper-connected, high-speed and mobile systems. In order to prepare for changes in the future, the Korean military intends to develop various weapons systems and form a multi-layer tactical network to support On The Move communication. However, current tactical networks are limited in support of On The Move communications. In other words, the operation of multi-beam antennas is necessary to efficiently construct a multi-layer tactical network in future warfare. Therefore, in this paper, we look at the need for multi-beam antennas through the operational scenario of a multi-layer tactical network. In addition, based on development consideration factors, features of rotary-wing and fixed-wing aircraft, we present the location and operation of airborne relay drone installations of multi-beam antennas.

전 세계적으로 사랑받고 있는 채소인 무는 뿌리채소 중 하나로 다양한 요리에 주재료로 사용되고 있어 많은 양이 생 산되고 소비된다. 무는 또한 한국의 음식 문화에서도 중요한 역할을 하는데, 대표적인 한국의 음식 중 하나인 김치 에도 무가 사용된다. 하지만 최근 급격한 기후 변화로 인해 무가 시들음병에 쉽게 노출될 수 있는 환경이 만들어져 무의 품질과 수확량이 크게 저하되는 문제가 생기고 있다. 무 시들음병 문제를 해결하기 위한 기존의 식물 대상의 질병 식별 방식은 수집한 컬러 이미지에서 수동으로 특징을 추출했기 때문에 많은 시간과 비용을 소모했다. 하지만 최근 근적외선 센서의 개발로 인해 식물의 질병 식별을 시간적, 금전적으로 보다 효율적으로 판별할 수 있도록 발전 하였다. 본 논문에서는 다중 이미지 센서를 기반으로 무인 비행체인 드론을 사용하여 무 시들음병을 식별할 수 있는 컬러 및 근적외선 이미지에 대한 딥 러닝 프레임워크를 제안하고 비교한다. 또한 사용자가 딥 러닝 모델의 결과를 시각적으로 이해할 수 있도록 설명가능한 인공지능(XAI) 접근 방식을 사용한다. 다양한 실험에서 얻은 결과로 제안 하는 프레임워크는 정확도, 계산 복잡성 측면에서 기존 탐지 시스템에 비해 더 나은 성능을 보인다는 것을 보여주었 고, 근적외선 데이터셋은 식생 지수 계산을 통해 무 시들음병을 식별하는데 효과가 있음을 보여주었다.

Radish, loved worldwide, is one of the root vegetables and is used as a main ingredient in various dishes, so it has a lot of production and consumption. Radish is also used in kimchi, one of the most popular Korean foods. Due to rapid climate change in recent years, it can be easily exposed to the radish fusarium wilt disease, which greatly reduces the quality and yield of radishes. Traditional plant disease identification methods used to solve the Fusarium Wilt problem are time-consuming and costly because they rely on manually extracting features from collected RGB images. Recent developments in near-infrared sensors have made it possible to determine plant health more efficiently, both in time and money. In this paper, we propose and compare Deep Learning Framework for RGB and NIR(Near-infrared) images that can identify radish fusarium wilt disease using Drone, Unmanned Aerial Vehicle based on multiple image sensors. It also uses an XAI(eXplainable Artificial Intelligence) approach that allows users to visually understand the results of the Deep Learning model. As a result of the various experiments, the proposed Framework showed better performance compared to existing detection systems in terms of accuracy and computational complexity, while NIR Dataset showed that it was effective in identifying the radish fusarium wilt disease through the Vegetation index calculation.

The paper proposed a method for controlling a robot arm mounted on an unmanned aerial vehicle(UAV). In this paper, the system is divided into a remote operating system and a control unit to perform tasks at a remote location by utilizing the characteristics of UAVs with no space limitations. For the experiment of the proposed multi-axis robot arm control system, a previously developed spider drone was used, and the multi-axis robot arm control system consists of a remote operating system for monitoring and giving commands, a controller for motion creation, and an image converter for transmitting camera images. We proposed a control method for a robot arm mounted on an UAV.

For multi- multi-unmanned aerial vehicle (UAV) air combat decision problems, this paper proposes a multi-UAV task allocation method based on cloud model. Firstly, we analyses air combat situation advantage of random and fuzzy information. The cloud model is applied to multi-UAV task allocation. We establish air combat superiority index model based on cloud model. And then through comprehensive cloud generation algorithm, we establish combat situation advantages of comprehensive cloud model based on cloud model and present multi-UAV task assignment method based on comprehensive cloud model. Finally, the simulation verification indicate: To evaluate air combat advantage using cloud model, we can get precise and credible combat situation advantage, thus, we get more scientific and reliable multi-UAV task assignment.

This paper presents an analysis method aiming at UAV cooperative detection problem. The UAV detect the target and establish the state information respectively, and send the message to the adjacent friendly aircraft. The friendly aircraft judge if there are repetition between their own information and the message they have received after they receive the message, and re-order the target till the state information of all UAV are same and complete the mission of multi-UAV distributed cooperative detection. The simulation shows that the algorithm presented in this paper could accomplish the cooperative detection mission effective and reasonable.

This paper presents advancement in multi- unmanned aerial vehicle (UAV) cooperative area surveillance, focusing on optimizing UAV route planning through the application of genetic algorithms. Addressing the complexities of comprehensive coverage, two real-time dynamic path planning methods are introduced, leveraging genetic algorithms to enhance surveillance efficiency while accounting for flight constraints. These methodologies adapt multi-UAV routes by encoding turning angles and employing coverage-driven fitness functions, facilitating real-time monitoring optimization. The paper introduces a novel path planning model for scenarios where UAVs navigate collaboratively without predetermined destinations during regional surveillance. Empirical evaluations confirm the effectiveness of the proposed methods, showcasing improved coverage and heightened efficiency in multi-UAV path planning. Furthermore, we introduce innovative optimization strategies, (Foresightedness and Multi-step) offering distinct trade-offs between solution quality and computational time. This research contributes innovative solutions to the intricate challenges of cooperative area surveillance, showcasing the transformative potential of genetic algorithms in multi-UAV technology. By enabling smarter route planning, these methods underscore the feasibility of more efficient, adaptable, and intelligent cooperative surveillance missions.

유연한 이동성, 쉬운 배치, 저렴한 비용 등의 장점을 가진 Unmanned Aerial Vehicle(UAV)를 이용해 Deep Neural Network(DNN) 서비스를 제공하는 기술이 연구되고 있다. 하지만 UAV 는 메모리와 컴퓨팅 능력, 배터리가 제한되어 있어 DNN 서비스의 요구사항을 만족시키기 위해서는 다수의 UAV간의 협업이 필요하다. 본 논문에서는 다수의 UAV 협업 환경에서 DNN 서비스의 처리 지연시간을 줄이기 위해 UAV 들의 작업량을 고려한 서비스 분산 처리 기법을 제안하고, 시뮬레이션을 통해 DNN 서비스 처리 지연 시간을 분석한다.

Coordination of multiple UAVs is an essential technology for various applications in robotics, automation, and artificial intelligence. In general, it includes 1) waypoints assignment and 2) trajectory generation. In this paper, we propose a new method for this problem. First, we modify the concept of the standard visibility graph to greatly improve the optimality of the generated trajectories and reduce the computational complexity. Second, we propose an efficient stochastic approach using simulated annealing that assigns waypoints to each UAV from the constructed visibility graph. Third, we describe a method to detect collision between two UAVs. FinallY, we suggest an efficient method of controlling the velocity of UAVs using A* algorithm in order to avoid inter-UAV collision. We present simulation results from various environments that verify the effectiveness of our approach.

Tilt rotor-wing concept to show enhanced performance in low speed mission is presented. Three types of stud wings on the existing tilt rotor configuration are suggested and their characteristics are compared. Aerodynamic analysis indicates that the stud wing concept gives significant performance improvement on the endurance and range in the low speed regime when compared with the tilt rotor. Penalties of the stud wing are discussed from the perspectives of conversion corridor, structural weight, configuration design, and cross wind stability. This study concludes that the advantage of the stud wing in general UAV mission performance is so significant as to surpass the penalties in other perspectives investigated.

Since the use of UAV (Unmanned Aerial Vehicle) is convenient for the acquisition of data on broad or inaccessible regions, it is nowadays used to establish spatial information for various fields, such as the environment, ecosystem, forest, or for military purposes. In this study, the process of estimating FVC (Fractional Vegetation Cover), based on multi-spectral UAV, to overcome the limitations of conventional methods is suggested. Hence, we propose that the FVC map is generated by using multi-spectral imaging. First, two types of result classifications were obtained based on RF (Random Forest) using RGB images and NDVI (Normalized Difference Vegetation Index) with RGB images. Then, the result map was reclassified into vegetation and non-vegetation. Finally, an FVC map-based RF were generated by using pixel calculation and FVC map-based GI (Gutman and Ignatov) model were indirectly made by fixed parameters. The method of adding NDVI shows a relatively higher accuracy compared to that of adding only RGB, and in particular, the GI model shows a lower RMSE (Root Mean Square Error) with 0.182 than RF. In this regard, the availability of the GI model which uses only the values of NDVI is higher than that of RF whose accuracy varies according to the results of classification. Our results showed that the GI mode ensures the quality of the FVC if the NDVI maintained at a uniform level. This can be easily achieved by using a UAV, which can provide vegetation data to improve the estimation of FVC.

Among various applications of the High Altitude Long Endurance (HALE) Unmanned Aerial Vehicle (UAV), this paper has a focus on the Global Positioning System (GPS) utilizing pseudolite and its improved performance, particularly during the multi-purpose missions. In a multi-purpose mission, the HALE UAV follows a specified flight trajectory for both navigation applications and missions. Some of the representative HALE missions are remote exploration, surveillance, reconnaissance, and communication relay. During these operations, the HALE UAV can also be an additional positioning signal source as it broadcast signals using pseudolite. The pseudolite signal can improve the availability, accuracy, and reliability of the GPS particularly in areas with poor signal reception, such as shadowed regions between tall buildings. The improvement in performance of navigation is validated through simulations of multi-purpose missions of the solar-powered HALE UAV in an urban canyon. The simulation includes UAV trajectory generation at stratosphere and uses actual geographical building data. The results indicate that the pseudolite-equipped HALE UAV has the potential to enhance the performance of the satellite navigation system in navigationally degraded regions even during multi-purpose operations.

This paper proposes a multi-UAV detection model using Continuous Wavelet Transform (CWT) and LeNet-5. Audio signals from UAVs are transformed via CWT to extract features, which are then classified using the LeNet-5 model. The dataset contains audio signals from eight UAV types, one unknown UAV class, and one for environmental noise - totaling ten classification labels. The proposed approach achieved over 95% accuracy in both training and real-world field tests. It is shown that detections and identifications of unknown drones based on audio signals are possible through the proposed method even under interfering sounds.

다양한 임무에서 활용 가능한 무인기 다개체 시스템은 단일 무인기보다 복잡하므로 효율적인 대형 제어방식이 요구된다. 특히 광역 탐색임무에 있어 통신량 및 연산량 부담이 적으며, 무인기간 자율적인 대형 형성이 가능한 분산 제어형의 유동적인 대형 형성이 필요하다. 본 연구는 스캔 면적의 확장 및 탐색 성능향상을 위해 Swarm 대형과 뱅크 정렬 대형, 대형 전체 운동을 고려한 대형 형성 알고리즘을 제안한다. 본 알고리즘은 상대거리에 대해서 2차 진동 특성을 가지며 parameter tuning을 통해 알고리즘을 설계할 수 있다. 또한 통상적인 무인기 시스템에 적합하도록 제어명령을 변환하였고, 시뮬레이션을 통해 알고리즘의 대형 형성 및 운동에 대한 성능을 입증하였다.

Since the Multi-UAV system for various missions is more complex than a single UAV, an efficient formation control method is required. In wide-area search mission, there is a need for a distributed control for flexible formation that has a low burden of communication and computation and enables autonomous formation between UAVs. This paper proposes a flexible formation operation method that considers the swarm formation, the bank alignment formation, and the formation movement to expand the scan area and improve search performance. The algorithm has a vibration characteristic of the second-order system for a relative distance and can design an algorithm through parameter tuning. In addition, we converted control commands to suit conventional UAV systems and demonstrated the performance of algorithms for a formation and movement of a formation through simulation.