http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Real-time 3D multi-pedestrian detection and tracking using 3D LiDAR point cloud for mobile robot
나기인,박병재 한국전자통신연구원 2023 ETRI Journal Vol.45 No.5
Mobile robots are used in modern life; however, object recognition is still insufficient to realize robot navigation in crowded environments. Mobile robots must rapidly and accurately recognize the movements and shapes of pedestrians to navigate safely in pedestrian-rich spaces. This study proposes real-time, accurate, three-dimensional (3D) multi-pedestrian detection and tracking using a 3D light detection and ranging (LiDAR) point cloud in crowded environments. The pedestrian detection quickly segments a sparse 3D point cloud into individual pedestrians using a lightweight convolutional autoencoder and connected-component algorithm. The multi-pedestrian tracking identifies the same pedestrians considering motion and appearance cues in continuing frames. In addition, it estimates pedestrians' dynamic movements with various patterns by adaptively mixing heterogeneous motion models. We evaluate the computational speed and accuracy of each module using the KITTI dataset. We demonstrate that our integrated system, which rapidly and accurately recognizes pedestrian movement and appearance using a sparse 3D LiDAR, is applicable for robot navigation in crowded spaces.
코파일럿 시스템을 위한 다중 2D LiDAR 융합기반 차량 주변 동적 장애물 추적 모듈 개발
나기인(Ki-in Na),변재민(Jaemin Byun),노명찬(Myoungchan Rho),서범수(Bumsu Seo) 한국자동차공학회 2014 한국자동차공학회 부문종합 학술대회 Vol.2014 No.5
This paper introduces the development of moving obstacles perception module using multiple 2D LiDARs. This can estimate both position and velocity of dynamic objects around vehicles. Projecting point cloud data from 2D LiDARs to range image structure, they are partitioned to particular observations with connectivity based region growing segmentation. In sequence, observations from segmentation steps are associated with the predicted tracks employing NNSF based on Kalman filter. Furthermore, the associated tracks are continuously updated, extraneous observations are generated to new tracks and missing tracks are removed. This developed module was installed on Co-Pilot system and the experiments tracking several moving obstacles at the same time were performed to show this module working.
자율 주행 차량의 레이저 기반 환경인식 모듈 설계 및 데이터 수집 시스템 구축
나기인(Ki-In Na),변재민(Jaemin Byun),노명찬(Myoungchan Roh),손주찬(Joochan Sohn),서범수(Bumsoo Seo),김성훈(Sunghoon Kim) 한국자동차공학회 2013 한국자동차공학회 부문종합 학술대회 Vol.2013 No.5
Architecture of the environmental perception module has been designed and specifically, the LiDAR based obstacle perception module has been designed and described with applicable algorithms. Moreover, the data logging system has been developed by vehicle, LiDAR and ROS packages for testing perception algorithms. The developed data logging system has been tested around ETRI and the logged data have been played back.
장지호,나기인,신호철,Chang, J.H.,Na, K.I.,Shin, H.C. 한국전자통신연구원 2022 전자통신동향분석 Vol.37 No.1
With the development of artificial intelligence, many studies have focused on evaluating abnormal situations by using various sensors, as industries try to automate some of the surveillance and security tasks traditionally performed by humans. In particular, mobile robots using multimodal sensors are being used for pilot operations aimed at helping security robots cope with various outdoor situations. Multiagent systems, which combine fixed and mobile systems, can provide more efficient coverage (than that provided by other systems), but network bottlenecks resulting from increased data processing and communication are encountered. In this report, we will examine recent trends in object recognition and abnormal-situation determination in various changing outdoor security robot environments, and describe an outdoor security robot platform that operates as a multiagent equipped with a multimodal sensor.
신호철,나기인,장지호,엄태영 한국전자통신연구원 2022 ETRI Journal Vol.44 No.2
Smart cities are expected to provide residents with convenience via various agents such as CCTV, delivery robots, security robots, and unmanned shuttles. Environmental data collected by various agents can be used for various purposes, including advertising and security monitoring. This study suggests a surveillance map data framework for efficient and integrated multimodal data representation from multi-agents. The suggested surveillance map is a multilayered global information grid, which is integrated from the multimodal data of each agent. To confirm this, we collected surveillance map data for 4 months, and the behavior patterns of humans and vehicles, distribution changes of elevation, and temperature were analyzed. Moreover, we represent an anomaly detection algorithm based on a surveillance map for security service. A two-stage anomaly detection algorithm for unusual situations was developed. With this, abnormal situations such as unusual crowds and pedestrians, vehicle movement, unusual objects, and temperature change were detected. Because the surveillance map enables efficient and integrated processing of large multimodal data from a multi-agent, the suggested data framework can be used for various applications in the smart city.
자율주행 시스템을 위한 다양한 도로 환경의 다중 LiDAR 센서를 이용한 데이터셋 구축 방법
변재민(Jaemin Byun),나기인(Ki-In Na),노명찬(Myoungchan Roh),서범수(Bumsoo Seo) 한국자동차공학회 2014 한국자동차공학회 부문종합 학술대회 Vol.2014 No.5
This paper describes a data set collected by multiple LiDAR sensors and describes how to build this data set for evaluation of perception algorithm in the intelligent vehicle. For covering the surrounding of vehicle, it contains data from 1 camera, two 4-layered LiDARs(LD-MRS) and a high density 3D LiDAR(Velodyne) were synchronized and saved to our logging system while our vehicle drives in a various types of road such as uphill, downhill, etc. In addition to, this data set also have ground truth labels such as road, vehicle, building ,etc. with corresponding to each 3D point. We wish that this data set will be useful to the autonomous vehicle community, especially those developing perception capabilities.
자율주행 자동차를 위한 3D LiDAR 정보를 이용한 MRF(Markov Random Field)기반 도로 인식 방법
변재민(Jaemin Byun),나기인(Ki-In Na),노명찬(Myoungchan Roh),손주찬(Joochan Sohn),서범수(Bumsoo Seo),김성훈(Sunghoon Kim) 한국자동차공학회 2013 한국자동차공학회 부문종합 학술대회 Vol.2013 No.5
본 논문은 자율주행자동차의 핵심 기술인 주행환경 인식기술과 관련하여 도로 영역인 주행 가능한 영역을 인식하는 방법에 대해서 제안한다. 기존 많이 사용했던 2D LiDAR 또는 카메라를 이용한 방법들은 2D 정보 기반으로 평지 환경에서의 주행 가능영역 인식에 대한 연구는 많이 진행되어 왔지만, 오르막(내리막) 또는 굴곡이 있는 도로 등에 대한 다양한 도로 유형에 대한 고려는 아직 미흡한 실정이다. 본 논문에서 굴곡이 있는 도로 환경에서 주행가능영역인식의 성능을 향상시키기 위해서 3D Lidar 센서로 획득된 3D Point Cloud 정보를 기반으로 기존의 널리 사용되었던 수직 방향의 높이정보를 통한 도로 영역을 분류하는 방법과 달리 Gradient 정보 기반으로 MRF 모델을 적용하고, LBP(Loopy Belief Propagation)방법을 통하여 도로 영역을 분류하는 방법을 제안하였다. 본 방법의 타당성을 보이기 위해서 실제 실험차량을 통해서 실제 도로의 데이터를 획득하고, 각도로 유형별에 따른 결과 및 분석을 통해서 본 방법의 성능의 우수성을 제시하였다.