Bimanual Manipulation and Multi-Agent Locomotion Tasks

该数据集包含了一系列在模拟环境中需要多个智能体协同行为的任务，特别关注双手操作和多智能体运动。具体包括开瓶、捡球、旋转瓶盖、拿起横杆、推蚂蚁和踢足球等多种任务，为评估强化学习中内在动机策略提供了多样化的场景。该数据集规模较大，拥有多个任务以及不同的配置。总体任务集中在双手操作和多智能体运动任务上。

This dataset contains a series of tasks requiring collaborative behaviors among multiple agents in simulated environments, with a particular focus on bimanual manipulation and multi-agent locomotion. It includes diverse tasks such as opening bottles, collecting balls, twisting bottle caps, lifting crossbars, pushing ants, and playing football, providing diversified scenarios for evaluating intrinsic motivation strategies in reinforcement learning. This large-scale dataset features multiple tasks and different configurations, with its overall task suite centered on bimanual manipulation and multi-agent locomotion tasks.