AdaR2R
收藏资源简介:
AdaR2R数据集是由新加坡理工大学创建的一个样本数据集,旨在支持在连续室内环境中进行视觉语言导航的研究。该数据集包含9个导航任务,每个任务在3个不同的Matterport3D场景中进行,涉及动态障碍物的路径规划。数据集的创建过程包括手动设置人类障碍物的生成点和路径,以增加导航任务的真实性。AdaR2R数据集主要应用于机器人导航和避障研究,旨在解决在复杂动态环境中进行有效导航的问题。
The AdaR2R dataset is a sample dataset created by Singapore Institute of Technology, designed to support research on vision-language navigation in continuous indoor environments. It contains 9 navigation tasks, each conducted in 3 distinct Matterport3D scenes, involving path planning with dynamic obstacles. The dataset creation process includes manually setting the spawn points and movement paths of human obstacles to enhance the realism of the navigation tasks. The AdaR2R dataset is primarily applied to research on robotic navigation and obstacle avoidance, aiming to address the challenge of effective navigation in complex dynamic environments.

- 1AdaVLN: Towards Visual Language Navigation in Continuous Indoor Environments with Moving Humans新加坡理工大学 · 2024年



