Robotic Task and Motion Planning Benchmark

该数据集包含了多种规划问题实例，这些实例以机器人任务和运动规划的障碍物布局及目标物体形状为特点。该数据集用于训练一个模型，以优化基于先前规划经验中的决策变量和得分对的评分函数。此外，该数据集源自成功的规划方案，并包含了诸如无碰撞运动和机器人操作任务中运动学上可行的分配等特征。其规模涵盖了过去规划经验中的多个问题实例，任务旨在优化机器人任务和运动规划的评分函数。

This dataset includes a variety of planning problem instances, which are defined by the obstacle layouts and target object shapes involved in robotic task and motion planning. It is utilized to train a model for optimizing the scoring function based on decision variables and score pairs from prior planning experiences. Additionally, this dataset is derived from successful planning solutions and encompasses features such as collision-free motions and kinematically feasible allocations in robotic manipulation tasks. It covers multiple problem instances from past planning experiences, with the core objective of optimizing the scoring function for robotic task and motion planning.