Earticle

This study develops a real-time ground coordinate acquisition system using a Laser Range Finder (LRF) mounted on an Unmanned Aerial Vehicle (UAV). To meet the requirements for precision and rapid response in surveillance, search and rescue, and disaster management, we implemented an integrated system comprising UAV hardware, coordinate calculation software, and a Ground Control Station (GCS) user interface. The system computes target locations in real-time by fusing sensor data from the LRF and UAV telemetry. Experimental results demonstrate that the proposed system maintains consistent coordinate accuracy within approximately 1.6m (mean) under static target conditions, across flight altitude variations of 10–50m and horizontal distance variations of 5–15m. This technology is expected to expand the operational scope of UAVs in both civilian and military sectors and enhance the technical competitiveness of domestic UAV-based positioning systems.

The Automatic License Plate Recognition (ALPR) is one of the intelligent transportation systems which provides a safe and secure mode of transportation. In ALPR technology, recognition accuracy entirely depends on the performance of the localization phase. This paper presents the real-time Bhutanese license plate (LP) localization using YOLO (You Only Look Once). The vehicle detection was performed before the LP localization to eliminate the false positives generated by the signboards as they look similar to LPs. A single convolutional neural network gave an overall mean average precision of 98.6% with the training loss of 0.0231 for vehicle and LP.

차량 측위 기술은 차량의 정확한 위치 정보를 제공한다는 점에서 자율주행을 위한 핵심 기 술 중 하나로 평가되고 있다. 이미지 기반의 측위 기술은 위치 정보를 효율적으로 제공할 수 있을 것으로 판단되어 다양한 관련 연구가 진행되고 있다. 하지만, 기존 특징점 또는 차선 정보 를 이용한 이미지 기반 측위 방법론은 도로 및 운행 환경에 측위 정확도가 큰 영향을 받을 수 있다는 한계가 있다. 선분 매칭은 특징점에 비하여 텍스쳐 반복에 강건하고 주변 환경 전체에 서 추출된 선분을 활용하기 때문에 기존 방법론의 단점을 해결할 수 있다. 하지만, 차량 운행 환경을 대상으로 한 선분 매칭 방법론을 다루는 연구는 거의 진행된 바 없다. 따라서 본 연구 에서는 정확한 차량 측위 지원을 위한 선분 매칭 프레임워크를 제안한다. 또한 도로 주행 환경 에서의 알고리즘 성능 비교 분석을 통하여 최적 선분 매칭 알고리즘을 결정하였다. 최종적으 로 제안된 프레임워크는 선분 추출, 병합, 중첩 영역 탐지 및 MSLD 기반 선분 매칭의 4단계로 구성되었다. 제안된 프레임워크는 차량의 속도, 운행 방식, 주변 환경에 상관없이 차량 측위에 충분한 수준의 선분 매칭을 안정적으로 수행하였다.

Vehicle localization is one of the core technologies for autonomous driving. Image-based localization provides location information efficiently, and various related studies have been conducted. However, the image-based localization methods using feature points or lane information has a limitation that positioning accuracy may be greatly affected by road and driving environments. In this study, we propose a line segment matching framework for accurate vehicle localization. The proposed framework consists of four steps: line segment extraction, merging, overlap area detection, and MSLD-based segment matching. The proposed framework stably performed line segment matching at a sufficient level for vehicle positioning regardless of vehicle speed, driving method, and surrounding environment.

In this paper, we present the design and implementation of real-time sound localization based on $0.13{\mu}m$ CMOS process. Time delay of arrival (TDOA) estimation was used to obtain the direction of the sound signal. The sound localization chip consists of four modules: data buffering, short-term energy calculation, cross correlation, and azimuth calculation. Our chip achieved real-time processing speed with full range ($360^{\circ}$) using three microphones. Additionally, we developed a dedicated sound localization circuit (DSLC) system for measuring the accuracy of the sound localization chip. The DSLC system revealed that our chip gave reasonably accurate results in an experiment that was carried out in a noisy and reverberant environment. In addition, the performance of our chip was compared with those of other chip designs.

Autonomous vehicle technology based on information technology and software will lead the automotive industry in the near future. Vehicle localization technology is a core expertise geared toward developing autonomous vehicles and will provide location information for control and decision. This paper proposes an effective vision-based localization technology to be applied to autonomous vehicles. In particular, the proposed technology makes use of numerical maps that are widely used in the field of geographic information systems and that have already been built in advance. Optimum vehicle ego-motion estimation and road marking feature extraction techniques are adopted and then combined by an extended Kalman filter and particle filter to make up the localization technology. The implementation results of this paper show remarkable results; namely, an 18 ms mean processing time and 10 cm location error. In addition, autonomous driving and parking are successfully completed with an unmanned vehicle within a $300m{\times}500m$ space.

Even if there are lots of researches on localization using 2D range finder in static environment, very few researches have been reported for robust real-time localization of mobile robot in uncertain and dynamic environment. In this paper, we present a new localization method based on ICP(Iterative Closest Point) algorithm for navigation of mobile robot under dynamic or uncertain environment. The ICP method is widely used for geometric alignment of three-dimensional models when an initial estimate of the relative pose is known. We use the method to align global map with 2D scanned data from range finder. The proposed algorithm accelerates the processing time by uniformly sampling the line fitted data from world map of mobile robot. A data filtering method is also used for threshold of occluded data from the range finder sensor. The effectiveness of the proposed method has been demonstrated through computer simulation and experiment in an office environment.

Indoor real-time positioning for multiple targets is required to realize human-robot symbiosis. This study firstly presents positioning accuracy on an autonomous mobile robot controlled by 3-D coordinates that is obtained by a real-time indoor positioning system with spread spectrum (SS) ultrasonic signals communicated by code-division multiple access. Although many positioning systems have been investigated, the positioning system with the SS ultrasonic signals can measure identified multiple 3-D positions in every 70 ms with noise tolerance and error within 100 mm. This system is also robust to occlusion and environmental changes. However, thus far, the positioning errors in an autonomous mobile robot, controlled by these systems using the SS ultrasonic signals, have not been evaluated as an experimental study. Therefore, a positioning experiment for trajectory control is conducted using an autonomous mobile robot and our positioning system. The effectiveness of this positioning method for robot self-localization is shown, from this experiment, because the average control error between the target position and the robot's position at 29 mm is obtained.

The most important factor of mobile robot is to build a map for surrounding environment and estimate its localization. This paper proposes a real-time localization and map building method through 3-D reconstruction using scale invariant feature from monocular camera. Mobile robot attached monocular camera looking wall extracts scale invariant features in each image using SIFT(Scale Invariant Feature Transform) as it follows wall. Matching is carried out by the extracted features and matching feature map that is transformed into absolute coordinates using 3-D reconstruction of point and geometrical analysis of surrounding environment build, and store it map database. After finished feature map building, the robot finds some points matched with previous feature map and find its pose by affine parameter in real time. Position error of the proposed method was maximum. 8cm and angle error was within $10^{\circ}$.

This study proposes a real-time target localization method based on acoustic refraction correction to enhance the accuracy of mine avoidance sonar systems for safe submarine operations. Conventional mine avoidance sonars typically assume either straight-line or single eigen-ray propagation paths for estimating target locations, which can result in detection errors due to insufficient consideration of acoustic refraction effects in complex and varying oceanic sound speed profiles. To address these limitations, this paper introduces a real-time sonar imaging method employing multi-ray tracing based on transmission angles and detailed sound speed profiles, effectively correcting mismatches in target localization. Performance validation through simulations and ocean experiments involving variable depths and complex sound speed environments demonstrated significantly improved localization accuracy compared to traditional approaches, confirming the capability for accurate real-time target localization in multi-target and complex maritime scenarios.

본 논문에서는 넓은 지역에서의 영상기반 자동 항법을 위한 실시간 위치인식 및 지도작성 방법을 제안한다. 한 대의 카메라에서 입력된 영상으로부터 제안된 방법은 6 자유도 카메라 자세와 3 차원 특징점 위치를 연속적으로 계산한다. 제안된 방법은 넓은 지역을 주행하며 촬영된 영상에 적용하여 그 위치와 환경지도를 성공적으로 작성하였이다. 본 논문에서는 이진기술자(binary descriptor)와 수치-위상(metric-topological)지도 표현법을 사용하여 GPU 나 영상의 축소 없이 실시간 성능과 광범위한 지역에서의 회귀점 검출(loop detection)을 하였다. 제안된 방법은 여러 환경에서 촬영된 영상과, 해당 영상의 GPS 기준값과 비교하여 평가하였다.

In this paper, we propose monocular visual simultaneous localization and mapping (SLAM) in the large-scale environment. The proposed method continuously computes the current 6-DoF camera pose and 3D landmarks position from video input. The proposed method successfully builds consistent maps from challenging outdoor sequences using a monocular camera as the only sensor. By using a binary descriptor and metric-topological mapping, the system demonstrates real-time performance on a large-scale outdoor dataset without utilizing GPUs or reducing input image size. The effectiveness of the proposed method is demonstrated on various challenging video sequences.

본 논문에서는 영상 감시 시스템을 위한 이동물체 추적 및 얼굴 영역 추출 알고리즘을 제안한다. 제안된 방법은 두 단계로 구성되어 있으며, 첫 번째의 이동 물체 감지 단계에서는 각각의 입력 프레임에 대하여 화소값의 변화를 기반으로 한 세 종류의 특징값들을 추출한 다음, 그 값들에 의하여 입력 프레임을 다섯 종류의 클래스로 분류한다. 분류 결과 현재의 프레임이 이동물체를 포함하고 있다면 두 번째의 얼굴 영역 분리 단계에서 컬러 정보와 프레임간의 차이를 이용하여 그 얼굴 영역을 추출하게 된다. 제안된 알고리즘은 광류등 계산시간이 많이 걸리는 특징값들에 의존하지 않으므로 실시간 처리에 적합하다. 또한 여러 테스트 영상시퀀스에 대하여 실험을 수행한 결과 본 논문에서 제안된 알고리즘의 유용성이 입증되었다.

In this paper, a moving object detection and face region extraction algorithm which can be used in video monitoring systems is presented. The proposed algorithm is composed of two stages. In the first stage, each frame of an input video sequence is analyzed using three measures which are based on image pixel difference. If the current frame contains moving objects, their skin regions are extracted using color and frame difference information in the second stage. Since the proposed algorithm does not rely on computationally expensive features like optical flow, it is well suited for real-time applications. Experimental results tested on various sequences have shown the robustness of the proposed algorithm.

This paper presents an implementation of real-time speaker localization using audio-visual information. Four channels of microphone signals are processed to detect vertical as well as horizontal speaker positions. At first short-time average magnitude difference function(AMDF) signals are used to determine whether the microphone signals are human voices or not. And then the orientation and distance information of the sound sources can be obtained through interaural time difference and interaual level differences. Finally visual information by a camera helps get finer tuning of the speaker orientation. Experimental results of the real-time localization system show that the performance improves to 99.6% compared to the rate of 88.8% when only the audio information is used.

본 논문에서는 실시간 스펙트럼 추정기에 기반하여 시간-주파수 영역 반사파 계측 시스템에서의 동축케이블 결함을 검출하는 방법을 제안한다. 시간-주파수 영역 반사파 계측 시스템은 가우시안 포락선(Gaussian Envelop)모양의 첩 신호를 기준신호로 하여 도선에 반사파를 분석하는 기법으로써 타 방법에 비하여 높은 정확도를 자랑하는 것으로 알려져 있다. 하지만 결함 위치를 추정하기 위해 Wigner 시간 주파수 분포 함수를 이용하므로 계산량 증가에 따른 실시간 구현이 어렵다는 단점이 있었다. 이러한 단점을 극복하기 위하여 본 논문에서는 LMS 스펙트럼 추정기를 이용한 시간-주파수 영역 반사파 계측 시스템의 구현방법을 새롭게 제안한다. 제안된 기법은 실제 동축케이블에 대한 실험결과를 통하여 그 성능을 입증하도록 한다.

Robots used for pipe inspection have been studied for a long time and many mobile mechanisms have been proposed to achieve inspection tasks within pipelines. Localization is an important factor for an inpipe robot to perform successful autonomous operation. However, sensors such as GPS and beacons cannot be used because of the unique characteristics of inpipe conditions. In this paper, an inpipe localization algorithm based on elbow detection is presented. By processing the projected marker images of laser pointers and the attitude and heading data from an IMU, the odometer module of the robot determines whether the robot is within a straight pipe or an elbow and minimizes the integration error in the orientation. In addition, an off-line positioning algorithm has been performed with forward and backward estimation and Procrustes analysis. The experimental environment has consisted of several straight pipes and elbows, and a map of the pipeline has been constructed as the result.

본 연구는 고정형 CCTV의 사각지대와 엣지 디바이스의 제한된 연산 자원 문제를 극복하기 위해, 실시간 재해 감지와 GPS가 없는 환경에서 자체 위치 추적이 가능한 경량화된 이동 로봇 시스템을 제안한다. 이를 위해 화재, 폭력, 작업자 자세 3가지 도메인의 지식을 단일 YOLOv8n 모델에 통합하는 다중 교사 지식 증류 기법을 적용하여 연산 효율성을 극대화하고, GPS 음영 지역에서의 정밀한 위치 추정을 위해 딥러닝 기반의 IMU 궤적 복원과 LiDAR, RGB-D 센서 융합 및 EKF(extended Kalman filter) 기반 위치 추정 아키텍처를 설계했다. NVIDIA Jetson Orin Nano 환경에서의 실증 실험 결과, 제안된 모델은 29 FPS(30.5ms)의 실시간 탐지 성능을 보여주며, 0.5m/s의 고속 주행 시에도 평균 0.67m의 위치 정확도를 유지하여 저사양 하드웨어 기반 시스템의 산업 현장 적용 가능성을 보여준다.

To overcome the blind spots of fixed CCTV systems and the computational constraints of edge devices, this paper proposes a lightweight mobile robot system capable of real-time industrial disaster detection and autonomous navigation. First, we implement a multi-teacher knowledge distillation framework that integrates heterogeneous domain knowledge ­ fire, violence, and pose - into a single YOLOv8n model to maximize computational efficiency. Second, for precise localization in GPS-denied environments, we design a sensor fusion architecture combining deep learning based IMU trajectory restoration with LiDAR and RGB-D sensors and Extended Kalman Filter. Experimental results using an NVIDIA Jetson Orin Nano demonstrate that the proposed system achieves a real-time detection speed of 30.5ms (29 FPS) and maintains a mean position error of 0.67m even during high-speed operation (0.5m/s). These findings validate the feasibility of deploying deep learning-based multi hazard detection and navigation on resource-constrained hardware in industrial settings.

영구자석 동기전동기 구동 인버터 시스템 내 스위치 고장은 다양한 조합으로 발생한다. 특히 스위치 개방고장의 경우 다른 전기 부품에 과전류를 유발하여 전체 시스템에 심각한 2차 손상을 초래한다. 본 논문에서는 개방고장 진단 알고리즘의 복잡성을 완화하기 위해 고장 그룹을 분류하여 고장 발생 시 false alarm의 영향을 받지 않고 간단하고 체계적인 방식으로 고장을 식별할 수 있는 기법을 제안한다. 스위치의 개방 고장으로 영구자석 동기전동기 구동 시스템 내 PWM 전압원 인버터의 신뢰성을 향상시키기 위해 3중 스위치 개방고장 진단 및 검출 기법을 제시한다. 제안된 기법은 기존의 개방고장 진단방법과 달리 정보처리를 위해 Moving Filter를 사용함으로써 연산부하를 증가시키지 않고 신속한 고장 검출이 가능하며 별도의 하드웨어 구성없이 구현이 가능하다. 제안된 기법의 안정성과 유효성이 시뮬레이션을 통해 입증된다.

Recent advances in outdoor autonomous robotics have highlighted the importance of precise localization using 3D point cloud-based mapping. Conventional full-map-based approaches suffer from scalability and latency issues in large-scale environments, making them less suitable for real-time applications. This paper proposes SRT-PM (Spacing-Radius-Time Partial Map), a method that partitions the environment into grids and selectively merges only the local regions around the robot in real time. By dynamically merging relevant point cloud segments based on the robot's position and movement, SRT-PM reduces computational overhead while maintaining localization accuracy. Experimental results demonstrate that the proposed method achieves up to 34% reduction in localization time and up to 98% reduction in point count compared to the full map approach (Large Area), with RMSE maintained or improved. This study contributes to enhancing real-time localization efficiency in large outdoor environments.