Earticle

차량 등 객체를 추적하기 위한 많은 알고리즘들이 있지만 본 논문에서 제안하는 특징점 정합 알고리즘 분야는 지수 복잡도의 시간이 걸리는 작업이다. 더구나, 차량을 추적하기 위해 기존에 제안되었던 객체 추출 등 영상전처리 알고리즘 또한 상당한 시간을 요구한다. 따라서 본 논문에서는 도로상에서 많은 차량들의 이동 궤적을 빠르고 효율적으로 추적하기 위한 방법을 2단계로 제안한다. 1단계로 객체 탐지가 아닌 번호판 영역을 먼저 탐지한 후 특징점을 추출하는 단계하고, 2단계로 특징점들을 정합하기 위한 비용산정식을 구한 후 동적계획법을 이용하여 효율적으로 차량을 추적할 수 있는 방법을 제안한다.

In the past, there have been many theory and algorithms for vehicle tracking. But the time complexity of many feature point matching methods for vehicle tracking are exponential. Also, object segmentation and detection algorithms presented for vehicle tracking are exhaustive and time consuming. Therefore, we present the fast and efficient two stages method that can efficiently track the many moving vehicles on the road. The first detects the vehicle plate regions and extracts the feature points of vehicle plates. The second associates the feature points between frames using dynamic programming.

Recently, RFID technology is being widely embraced in the supply chain, by manufacturers, retailers, and logistics firms. This study shows the RFID system to provide location based service for commercial vehicle by integrating tracking function of GPS with Mobile service. For this, we configured PC-based terminal for vehicle and developed application consisted of RFID and GPS manager program. In web service module which uses mobile service, commodity information acquired from each vehicle at real-time is added at database and it is analyzed in various aspects to support management decision through the Legacy system.

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. 2018 The Korean Institute of Communications and Information Sciences. Publishing Services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

본 연구에서는 타이어를 장착한 차량시험기에서 교량과 차량의 상호작용효과를 분석하고자 반복하중시험을 실시하고 이에 대한 동적반응을 파악하였다. 이를 위하여 차량-교량간 상호작용이 포함된 이동질량형 윤하중 실험기를 단순교 형식의 교량에 적용하여 이동질량 반복주행실험을 수행하였다. 이동질량 반복주행실험 결과를 분석하여 차량-교량간의 상호작용을 포함한 가속도 형태의 실동주행차량가속도를 규명하였다. 규명된 차량의 실동주행차량가속도를 범용 해석프로그램에 적용하여 차량-교량 상호작용을 재현할 수 있었다.

In this paper, an analytical and experimental study is performed in order to determine the effects of interaction between a vehicle and a structure. For this purpose, a wheel tracking machine and an adequate single span bridge are designed. Results presented in the paper show that the real vehicle tracking accelerations including the interaction between the vehicle and the structure produce additional effects on the dynamic behavior of the structure including reversal and contrary behavior. Also, the interaction between the vehicle and the bridge is reproduced by applying the identified real vehicle tracking accelerations to a general finite element analysis program.

비디오를 이용한 차량 추적 자동화 기술은 차량속도 등의 정확한 교통량 측정을 위해 중요하다. 그러나 단일 카메라로 취득된 비디오를 이용한 차량추적의 경우 카메라의 유한한 시계(field of view)때문에 추적범위가 제약된다. 차량추적 범위를 넓히고 차량행태에 대한 보다 많은 정보를 얻기 위해서는 복수개의 카메라로 연속된 추적영역을 촬영하여 분석하는 방식을 이용할 수 있다. 본 논문에서는 규칙기반 차량추적 알고리즘을 적용한 두 대의 동기된 비디오 시??스에서 연속적으로 추적하는 방식을 제안한다. 제안된 방법에서는 영상분석을 통해 나타날 수 있는 거의 모든 경우를 규칙을 만들기 위해 고려되었으며, 칼만 필터를 이용하여 예측성능을 향상시켰다. 제안된 연속된 비디오에서의 차량추적 방법은 확장된 추적범위에서 차량의 순간속도, 차량의 개별행태 등에 보다 많은 정보를 취득하는데 성공적으로 활용될 수 있다.

Automatic tracking of vehicles is important to accurately estimate the traffic information including vehicle speeds in video-based traffic measurement systems. Because of the limited field of view, the range of visual tracking with a single camera is restricted. In order to enlarge the tracking range for better chance of monitoring the vehicle behaviors, a tracking with consecutive multiple video sequences is necessary. This parer proposes a carefully designed rule-based vehicle racking scheme and apply it for the tracking for two well synchronized video sequences. In the scheme, almost all possible cases that can appear in the video-based vehicle tracking are considered to make rules. Also, the rule based scheme is augmented with Kalman filter. The result of tracking can be successfully used to collect data such as temporal variation of vehicle speed and behavior of individual vehicle behaviors in the enlarged tracking region.

차량의 자동 추적은 도로 건설 계획을 수립하는데 있어서 중요한 자료인 교통의 흐름을 분석하거나, 비디오 영상의 교통량 분석 시스템에서 차량의 정확한 속도를 측정하는 등 여러 분야에서 그 중요성이 날로 증가하고 있다. 본 논문에서는 비디오 영상에서 차량 추적 시 발생할 수 있는 가능한 경우들을 고려한 규칙기반 차량 추적 방법들을 제안한다. 제안된 방법들은 정화도의 관점에서 Mean-Shift방법에 비해 빠르고 좋은 성능을 나타내었다 본 논문에서 제안된 방법은 추적의 속도나 정확성을 높이기 위해 칼만 필터 예측기 및 칼라 정보를 이용한 탐색들을 융합한다면 더 우수한 추적 성능을 보일 것으로 기대된다.

Automatic tracking of vehicles is important to accurately estimate the vehicle speeds in video-based traffic measurement systems and to analyze traffic flows for road construction. This paper proposes a carefully designed rule-based tracking scheme that considers the possible cases that can be appeared in the video-based vehicle racking. The proposed scheme is fast and outperforms the Mean-Shift scheme in terms of accuracy. The accuracy and the speed of the scheme would be increased by combining it with color-based searching and Kalman filters.

There are many inductive loop detectors under the highways in Korea. Among the other detectors, some are image detectors. Almost all image detectors are focused one or two lane of the road and are measuring traffic information. This paper propOses to an algorithm for detecting traffic information automatically from CCTV camera images installed on the highway. The information which is counted in one lane or two contains some critical errors by occlusion frequently in case of passing larger vehicles. In this paper, we use a tracking algorithm in which the detection area include all lanes, then the traffic informations are collected from the vehicles individually using difference images in this detection area. This tracking algorithm is better than lane by lane detecting algorithm. The experiment have been conducted two different real road scenes for 20 minutes. For the experiments, the images are provided with CCTV camera which was installed at Kiheung Interchange upstream of Kyongbu highway, and video recording images at Chungkye Tunnel. For image processing, images captured by frame-grabber board 30 frames per second, 640×480 pixels resolution and 256 gray-levels to reduce the total amount of data to be interpreted.

본 논문에서는 확장형 칼만필터를 적용하여 영상센서 기반 차대도로간 트래킹 알고리즘 을 제안한다. 일반적으로 횡방향 오프셋, 차선대비 상대경로각, 전방도로 곡률은 차선유지지 원시스템의 경로추종 횡방향 제어기 구성 또는 차선이탈경보시스템의 경보 로직을 위한 중 요 입력값으로 활용되는데 이를 위해 본 연구에서는 영상센서 차선인식 결과값인 이미지 상 의 차선 추출점의 좌표값과 더불어 요레이트, 조향각, 차속 센서 측정값, 그리고 차량의 횡방 향 운동방정식을 고려한 확장형 칼만필터를 적용하여 차대도로간 트래킹 정보를 추출한다. 제안된 차대도로간 트래킹 알고리즘의 유효성을 증명하기 위해 주행 테스트 도로 상에서 DGPS-RTK 장비를 이용하여 비교 검증하여 그 유효성을 보였다.

This paper proposes a vehicle to road tracking algorithm based on vision sensor by using EKF(Extended Kalman Filter). The lateral offset, heading angle, and curvature which are obtained from vehicle to road tracking might be used as inputs to steering controller of LKAS(Lane Keeping Assist System) or for the warning decision logic of LDWS(Lane Departure Warning System). To the end, in this paper, the yaw rate, steering angle, and vehicle speed as well as lane raw points together with considering of vehicle lateral dynamics are utilized to improve the exactness and convergence of the vehicle to road tracking. The proposed algorithm has been tested at a proving ground that consists of straight and curve sections and compared with GPS datum using DGPS-RTK equipment to show the feasibility of the proposed algorithm.

본 논문에서는 가우시안 정규분포 모델 기반의 멀티센서 데이터 퓨전 알고리즘을 이용한 차량 위치인식 시스템을 제안한다. 기존의 차량 위치인식 시스템은 GPS를 중심으로 제공되어 왔으나 위성 신호 수신이 어려운 실내나 빌딩이 빽빽하게 들어선 도심에서는 제대로 작동하지 않는 문제가 있었다. 이를 해결하기 위해 GPS와 UWB를 결합하는 방법이 연구되었으나 위치변화에 따라 각 측위매체를 이산적으로 전환하여 차량과 같은 이동이 잦은 대상에게 끊임없이 유연한 위치 정보를 제공하지 못하는 단점이 있다. 따라서 본 논문에서는 가우시안 정규분포 모델을 바탕으로 GPS 위치 데이터와 UWB 센서 데이터를 유기적으로 결합하는 Hybrid UWB/GPS 측위시스템을 구성하여 측위시스템의 적용범위에 민감하지 않고 유연한 위치정보를 제공하도록 한다. 제안된 시스템을 Ubisense와 Asen GPS를 이용하여 12m × 8m 크기의 실외 환경에서 실험하였으며, 구현된 시스템을 통해 기존 UWB 및 GPS 시스템에 비해 정밀도 및 연동성이 향상되었음을 확인하였다. 이를 통해 차량에서 주차 관리 및 차량추적 관리 등 다양한 CNS를 지원할 수 있다.

In this paper, Gaussian probability distribution model based multi-sensor data fusion algorithm is proposed for a vehicular location awareness system. Conventional vehicular location awareness systems are operated by GPS (Global Positioning System). However, the conventional system is not working in the indoor of building or urban area where the receiver is difficult to receive the signal from satellites. A method which is combined GPS and UWB (Ultra Wide-Band) has developed to improve this problem. However, vehicular is difficult to receive seamless location information since the measurement systems by both GPS and UWB convert the vehicle’s movement information separately at each sensor. In this paper, normalized probability distribution model based Hybrid UWB/GPS is proposed by utilizing GPS location data and UWB sensor data. Therefore the proposed system provides information with seamless and location flexible properties. The proposed system tested by Ubisense and Asen GPS in the 12m×8m outdoor environments. As a result, the proposed system has improved performance for accurateness and connection ability between devices to support various CNS (Car Navigation System).

자율주행차량의 기술 발전과 보급률 증가로 인해 기존 비자율주행차량 교통사고와는 다른 유형의 사고가 새롭게 등장하고 있다. 본 연구는 다양한 센서를 기반으로 주행하는 자율주행 차량의 교통사고 원인을 분석하기 위해, 은닉 마르크프 모델(HMM)을 활용한 사고원인 추적 방법을 제안한다. 제안된 모델은 전체 사고 발생과정을 충돌 전(Pre-crash), 충돌 상태(In-crash), 충돌 후(Post-crash)로 구분하고, 각 단계의 상태 전이와 주요 요인 간 상호작용을 확률적 접근 법으로 분석할 수 있도록 설계되었다. 또한 3W (Who, Why, What) 포맷을 기반으로 주요 자율 주행교통사고 시나리오를 제시하여 모델의 활용 가능성을 검증하였다. 본 모델은 복잡한 사고 원인을 체계적으로 분석하는 데 높은 성능을 보이며, 특히 Pre-crash 단계에서 알 수 없는 블랙 데이터인 은닉 상태를 효과적으로 추적하여 사고 대응 및 예방에 기여할 것으로 기대된다.

With the advancement of autonomous vehicle technology and its increasing adoption, new types of accidents distinct from those involving conventional vehicles have emerged. This study proposes a Hidden Markov Model (HMM)-based approach to systematically track and analyze the causes of autonomous vehicle accidents. The proposed model categorizes the accident process into three stages: Pre-crash, In-crash, and Post-crash. It is designed to analyze the transitions between states in each stage and the interactions among key factors using a probabilistic approach. Furthermore, the study introduces practical applications of the model by presenting major autonomous vehicle accident scenarios based on the 3W (Who, Why, What) framework. The model demonstrates high performance in systematically analyzing complex accident causes, particularly by effectively tracking hidden states, or "black data," in the Pre-crash stage. This capability is expected to contribute significantly to improving accident response and prevention.

Vehicle detection and tracking applications play an important role for military and civilian applications such as in highway traffic surveillance control management and traffic planning.This paper presents a review on the various techniques of On-Road Vehicle detection systems that are based on motion model. In this paper a literature Survey of previous and recent works is presented on vision-based vehicle detection using sensors. Detecting the objects in the video and tracking their motion to identify their characteristics has been emerging as a demanding research area in the domain of Image Processing and Computer Vision. The traffic image analysis comprises of three parts: (1) Traffic Analysis (2) Motion Vehicle Detection and Segmentation Approaches and (3) Vehicle Tracking Approaches. In this survey, we have classified these methods into various groups, and these groups are providing a detailed description of various representation methods and find out their positive and negative aspects.

This paper provides a comprehensive overview of UAV classification, tracking, and detection, offering researchers a clear understanding of these fundamental concepts. It elucidates how classification categorizes UAVs based on attributes, how tracking monitors real-time positions, and how detection identifies UAV presence. The interconnectedness of these aspects is highlighted, with detection enhancing tracking and classification aiding in anomaly identification. Moreover, the paper emphasizes the relevance of simulations in the context of drones and UAVs, underscoring their pivotal role in training, testing, and research. By succinctly presenting these core concepts and their practical implications, the paper equips researchers with a solid foundation to comprehend and explore the complexities of UAV operations and the role of simulations in advancing this dynamic field.

In this paper, we propose a design method of the tracking and controlling AUV(autonomous underwater vehicle) that has wide range of velocity. The state feedback controllers are designed according to determined surge speed and then a fuzzy controller is designed in a way that the outputs of the state feedback controller are combined by the surge speed of AUV, so that when the surge speed changes, the AUV control becomes stable. By simulation using Matlab/Simulink, the performance of the proposed controller is shown to be efficient.