遇见数据集

临界场景数据与预期功能安全场景库第三类未知危险场景数据

收藏
官方服务:

资源简介:

测试场景数据集包括China-FOT自然驾驶数据和预期功能安全第三类未知场景场地测试数据。 China-FOT数据包括自然驾驶原始数据、自然驾驶行为统计数据以及临界场景特征数据。基于具有中国驾驶员行为特性的China-FOT自然驾驶数据,分析跟车、变道、切入等场景的特征要素,参考ECE R157标准优化驾驶员分级危险处置模型,为功能场景-逻辑场景-临界场景的测试用例筛选机制提供支撑。 预期功能安全场景库第三类未知危险场景包括14类测试场景,覆盖汽车、摩托车、自行车、成人假人、儿童假人、交通锥桶6类目标物,并分为异态目标物测试和雨雾模拟测试,包括车道内遭遇静止车辆、车道内遭遇摩托车、车道内遭遇自行车、车道内遭遇成人假人、车道内遭遇儿童假人、车道内遭遇成人及儿童假人、车道内遭遇锥桶、前车切出遭遇静止车辆、前车切出遭遇摩托车、前车切出遭遇自行车、前车切出遭遇成人假人、前车切出遭遇儿童假人、前车切出遭遇成人及儿童假人、前车切出遭遇锥桶等测试场景,共开展了149例封闭场地实车测试,包括87例异态目标物测试和61例雨雾模拟测试。测试过程通过外置摄像机、车内摄像头进行试验记录,测试数据包括车辆速度、加速度、纵向间距等试验数据。 数据集共有57个文件夹,466个文件,包括excel、mat、txt、jpg、fig、mp4、avi等格式的数据,数据量规模为2GB。

The test scenario dataset includes China-FOT natural driving data and third-category unknown scenario field test data for automotive intended functional safety. The China-FOT dataset includes raw natural driving data, statistical data on natural driving behaviors, and critical scenario feature data. Based on the China-FOT natural driving data that reflects the behavior characteristics of Chinese drivers, we analyze the characteristic elements of scenarios such as car-following, lane change, and cut-in, optimize the hierarchical driver hazard disposal model with reference to the ECE R157 standard, providing support for the test case screening mechanism of functional scenario, logical scenario, and critical scenario. The third-category unknown hazardous scenarios in the intended functional safety scenario library include 14 types of test scenarios, covering 6 types of objects: motor vehicles, motorcycles, bicycles, adult dummies, child dummies, and traffic cones. These scenarios are divided into abnormal object tests and rain/fog simulation tests. The specific scenarios include encountering stationary vehicles in the lane, encountering motorcycles in the lane, encountering bicycles in the lane, encountering adult dummies in the lane, encountering child dummies in the lane, encountering both adult and child dummies in the lane, encountering traffic cones in the lane, encountering stationary vehicles when the preceding vehicle cuts out, encountering motorcycles when the preceding vehicle cuts out, encountering bicycles when the preceding vehicle cuts out, encountering adult dummies when the preceding vehicle cuts out, encountering child dummies when the preceding vehicle cuts out, encountering both adult and child dummies when the preceding vehicle cuts out, and encountering traffic cones when the preceding vehicle cuts out. A total of 149 closed-field real-vehicle tests have been conducted, including 87 abnormal object tests and 61 rain/fog simulation tests. The test process is recorded by external cameras and in-vehicle cameras, and the test data includes experimental data such as vehicle speed, acceleration, and longitudinal spacing. The dataset contains a total of 57 folders and 466 files, including data in formats such as excel, mat, txt, jpg, fig, mp4, and avi, with a total data size of 2GB.

提供机构:
同济大学
搜集汇总
数据集介绍
临界场景数据与预期功能安全场景库第三类未知危险场景数据 数据集图片
背景与挑战
背景概述
该数据集包含China-FOT自然驾驶数据和预期功能安全第三类未知危险场景的场地测试数据,用于分析驾驶行为特征和支撑测试用例筛选。其中,场地测试覆盖了14类场景和6类目标物,共进行了149例实车测试,数据量约为2GB,包括多种格式的文件。
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务