Safety Gymnasium

该数据集是一个基准测试套件，旨在评估在高维复杂场景中的性能和安全，这些场景融合了多种安全约束，涉及使用激光雷达传感器的机器人任务。数据集包括半随机目标、随机目标以及随机圆圈等任务，每个任务都设计有不同的复杂性和安全约束。该数据集覆盖了高维复杂环境下的多项任务，其目标是评估强化学习代理在确保安全的同时，在障碍物环境中导航的性能。

This dataset is a benchmark suite developed to evaluate both performance and safety in high-dimensional complex scenarios that incorporate multiple safety constraints, targeting robotic tasks equipped with LiDAR sensors. The dataset includes a variety of tasks such as semi-random goal, random goal, and random circular obstacle scenarios, where each task is engineered with distinct complexity levels and safety constraints. Covering diverse tasks across high-dimensional complex environments, this benchmark aims to assess the performance of reinforcement learning agents when navigating safely in obstacle-cluttered environments.