NavigationNet

NavigationNet是由上海交通大学开发的大型室内导航数据集，旨在通过深度强化学习解决室内场景中的导航问题。该数据集覆盖约1500平方米的室内区域，包含多种房间类型，如卧室、会议室等。数据集的创建过程中，使用配备8个摄像头的机器人，在可行走位置捕捉图像，以模拟真实世界的导航环境。NavigationNet不仅支持机器人自主导航的训练，还为室内场景理解提供了丰富的视觉数据。该数据集的应用领域广泛，包括机器人导航、室内地图构建和智能空间管理等，旨在提高机器人在未知环境中的自主导航能力。

NavigationNet is a large-scale indoor navigation dataset developed by Shanghai Jiao Tong University, which aims to solve indoor navigation problems via deep reinforcement learning. This dataset spans an indoor area of approximately 1,500 square meters and encompasses diverse room types including bedrooms, meeting rooms, and more. During its construction, a robot equipped with 8 cameras was deployed to capture images at traversable positions, simulating real-world navigation environments. NavigationNet not only supports the training of robotic autonomous navigation systems but also provides abundant visual data for indoor scene understanding. It has broad application prospects in fields such as robotic navigation, indoor map construction, and intelligent space management, with the goal of improving the autonomous navigation capability of robots in unknown environments.