基于多源信息危险驾驶预警标注数据

基于开放智能网联车实验数据（SPMD，Safety Pilot Model Deployment），通过数据预处理、人工校验等方法，提取自然驾驶数据中危险驾驶事件，同时将危险驾驶事件同路网结构、天气环境数据进行合并与标注。构建了用于开展基于多源信息危险驾驶预警模型研究的数据集。该成果以自然驾驶数据为研究对象，应用机器学习技术对多源信息和危险驾驶事件进行建模，实验表明该模型预警性能优于单车动力学模型。数据类型为csv，记录数12202条数据。

Based on the open intelligent connected vehicle experimental dataset (SPMD, Safety Pilot Model Deployment), dangerous driving events were extracted from natural driving data through data preprocessing, manual verification and other processing steps. Subsequently, the dangerous driving events were merged and annotated with road network structure and weather environment data, thereby constructing a dataset for research on multi-source information-based dangerous driving early warning models. This work takes natural driving data as the research subject, applies machine learning technologies to model multi-source information and dangerous driving events, and experimental results demonstrate that the early warning performance of the proposed model outperforms that of the single-vehicle dynamics model. The dataset is in CSV format, containing a total of 12202 records.

搜集汇总

背景与挑战

背景概述

该数据集基于开放智能网联车实验数据，通过预处理和人工校验提取了12202条危险驾驶事件记录，并结合路网结构和天气环境数据进行标注，用于危险驾驶预警模型研究。数据格式为csv，总大小为2.34MB，包含2个文件。