MIMo Environment for Pulling-to-Stand Task

该数据集旨在通过在婴儿床中放置MIMo来学习使用本体感觉和前庭输入的拉起站立行为。在设计奖励函数时，我们优先考虑站立姿势，并对过大的力量施加惩罚。该任务是在单一代理的规模上进行的，任务是拉起站立。

This dataset aims to enable the learning of pull-to-stand behavior using proprioceptive and vestibular inputs, by placing MIMo in a crib. When designing the reward function, we prioritize the standing posture and impose penalties for excessive force. The task is conducted on a single-agent scale, with the core objective being pull-to-stand.