Wheel-Mounted Inertial Datasets

A wheel-mounted inertial sensor mitigates inertial drift more effectively than an inertial sensor mounted on the vehicle chassis. Although their usage is increasing, there is no publicly available dataset for wheel-mounted inertial sensors. To fill this gap, this work presents the wheeled-mounted inertial (WMI) dataset. WMI was recorded using two platforms: an omni-directional robot equipped with 5 IMUs, and a passenger car equipped with 9 IMUs. Each platform features IMUs mounted on every wheel. In total 106 minutes of recordings for each IMU (740 minutes for all IMUs) were made with associated ground truth trajectory. This versatile dataset will help develop model-based and data-driven approaches with wheel mounted inertial sensors.

相较于安装于车辆底盘的惯性传感器，轮载式惯性传感器（wheel-mounted inertial sensor）可更有效地抑制惯性漂移。尽管其应用场景不断拓展，但目前尚无针对轮载式惯性传感器的公开可用数据集。为填补这一研究空白，本研究特此推出轮载惯性（WMI）数据集。该数据集通过两类平台采集：一是搭载5个惯性测量单元（Inertial Measurement Unit，IMU）的全向移动机器人，二是搭载9个惯性测量单元的乘用车。两类平台的所有车轮均安装有惯性测量单元。每个惯性测量单元的采集时长总计为106分钟，所有惯性测量单元的总采集时长达740分钟，且同步配套有真值轨迹（ground truth trajectory）数据。该通用数据集将助力基于模型与数据驱动的轮载式惯性传感器相关研究方法的开发。