http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
열악한 환경에서의 자율주행을 위한 다중센서 데이터셋 구축
심성대,민지홍,안성용,이종우,이정석,배광탁,김병준,서준원,최선덕 한국로봇학회 2022 로봇학회 논문지 Vol.17 No.3
Sensor dataset for autonomous driving is one of the essential components as the deep learning approaches are widely used. However, most driving datasets are focused on typical environments such as sunny or cloudy. In addition, most datasets deal with color images and lidar. In this paper, we propose a driving dataset with multi-spectral images and lidar in adverse weather conditions such as snowy, rainy, smoky, and dusty. The proposed data acquisition system has 4 types of cameras (color, near-infrared, shortwave, thermal), 1 lidar, 2 radars, and a navigation sensor. Our dataset is the first dataset that handles multi-spectral cameras in adverse weather conditions. The Proposed dataset is annotated as 2D semantic labels, 3D semantic labels, and 2D/3D bounding boxes. Many tasks are available on our dataset, for example, object detection and driveable region detection. We also present some experimental results on the adverse weather dataset.
임의 주행 기반 3차원 라이다 좌표와 차량 차표의 외부 파라미터 캘리브레이션 기법
심성대(Sungdae Sim),민지홍(Jihong Min),김준(Jun Kim) 대한기계학회 2016 대한기계학회 춘추학술대회 Vol.2016 No.12
In many autonomous vehicles, 3D LADARs have become a very important sensor. In many case, we put it on someplace of vehicle such as the roof. To integrate sensor informations from time to time, we should know the sensor’s movement. In many applications, we know the movement from the vehicle’s movement via navigation systems such as GPS/IMU or vehicle odometers. When we know the rotations and translations between 3D LADAR and vehicle coordinates, the movement of the sensor can be obtain accurately via a simple transformation. We estimate the extrinsic parameters using iterative closest point algorithm and nonlinear optimization. Experimental results show the proposed algorithm works well.
심성대(Sungdae Sim),민지홍(Jihong Min),김병준(Byungjun Kim),김준모(Junmo Kim) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.11
Trajectory prediction is one of the important technology for autonomous vehicles and robots. Recently, many learning-based trajectory prediction algorithms are released. Most approaches tend to use recurrent neural networks such as LSTM. Our approach is trajectory prediction using attention mechanisms. Our experiments show the attention mechanism outperforms RNN.