NeurIT Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/NeurIT-Project/NeurIT
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是通过一个定制机器人平台在三栋具有不同布局、环境和地面条件的建筑中构建的，其中包含了用于评估TF-BRT模型有效性的追踪数据。数据集被分为四部分：训练集、验证集、测试可见集和测试未见集，其比例分别为15:3:3:4。该数据集包含了110个序列，总行程距离约为33.7公里，其任务是室内机器人惯性追踪。
This dataset was constructed using a custom robotic platform across three buildings with distinct layouts, environments, and ground conditions, and contains tracking data for evaluating the effectiveness of the TF-BRT model. The dataset is divided into four subsets: training set, validation set, test seen set, and test unseen set, with a split ratio of 15:3:3:4 respectively. It includes 110 sequences with a total travel distance of approximately 33.7 kilometers, targeting the task of indoor robotic inertial tracking.
提供机构：
NeurIT Project
