HDD

我们介绍了HRI驾驶数据集 (HDD)，这是一个具有挑战性的数据集，可以研究在现实生活环境中学习驾驶员行为。该数据集包括使用配备有不同传感器的仪表车辆在旧金山湾区收集的104小时的真实人类驾驶。下面的视频提供了数据集中面向目标的驾驶行为的示例。GPS坐标以及来自CAN总线的传感器值显示在前置摄像头流的顶部。

We introduce the HRI Driving Dataset (HDD), a challenging dataset for investigating driver behavior learning in real-world environments. This dataset comprises 104 hours of real human driving data collected in the San Francisco Bay Area using instrumented vehicles equipped with various sensors. The following video showcases examples of goal-oriented driving behaviors from the dataset. GPS coordinates and sensor values from the CAN bus are overlaid on the top of the front-facing camera stream.