http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Deep Neural Network 기반의 차량 주행 경로 예측
정동기(Donggi Jeong),백민진(Minjin Baek),김우중(Woojoong Kim),이상선(Sang-Sun Lee) 한국자동차공학회 2018 한국 자동차공학회논문집 Vol.26 No.2
In regard to the vehicle safety services, vehicle trajectory prediction is one of the key technologies to estimate when and where the vehicle could be located. The development of the vehicle trajectory prediction algorithms using traditional physics-based models, such as kinematic and dynamic models, often required building a system model with complex equations, as well as gathering sensor noise statistics. In this paper, we proposed a vehicle trajectory prediction algorithm based on a deep neural network(DNN). The input data to the DNN were the driving status information defined in SAE J2735 basic safety message(BSM), while the output layer consisted of longitudinal and lateral trajectory predictions. The adaptive moment estimation(Adam) was used to enhance the learning speed. Through simulations, we compared the performance of the DNN-based approach with that of the traditional approach, and the results showed that the proposed method resulted in improved accuracy and precision of the vehicle trajectory prediction.