RoboCupSimData

该数据集包含了2016年和2017年RoboCup足球仿真联赛（2D）顶级队伍的比赛数据，不仅包括每个机器人的真实数据，还有带噪声的局部数据。数据来源于10支队伍，涵盖了45组独特的对阵组合，同时为每个机器人提供了真实数据和局部感知数据。数据规模宏大，包含1125场比赛，超过180小时的游戏时长，压缩后为17GB（解压后为229GB）。这项数据的任务是对多机器人系统中机器学习的分析和应用研究。

This dataset contains match records from the top-tier teams of the RoboCup Soccer Simulation League (2D) in 2016 and 2017. It includes both the ground-truth data of each individual robot and noisy local perception data. Sourced from 10 participating teams, the dataset covers 45 unique matchups, and provides paired ground-truth and local perception data for every robot. The dataset has a considerable scale, consisting of 1125 matches totaling over 180 hours of gameplay, with a compressed size of 17 GB (229 GB when decompressed). This resource is intended for research on machine learning analysis and applications in multi-robot systems.