Daniel Butterfield, Sandilya Sai Garimella, Nai-Jen Cheng, Lu Gan
{"title":"MI-HGNN:用于腿部机器人接触感知的形态信息异构图神经网络","authors":"Daniel Butterfield, Sandilya Sai Garimella, Nai-Jen Cheng, Lu Gan","doi":"arxiv-2409.11146","DOIUrl":null,"url":null,"abstract":"We present a Morphology-Informed Heterogeneous Graph Neural Network (MI-HGNN)\nfor learning-based contact perception. The architecture and connectivity of the\nMI-HGNN are constructed from the robot morphology, in which nodes and edges are\nrobot joints and links, respectively. By incorporating the morphology-informed\nconstraints into a neural network, we improve a learning-based approach using\nmodel-based knowledge. We apply the proposed MI-HGNN to two contact perception\nproblems, and conduct extensive experiments using both real-world and simulated\ndata collected using two quadruped robots. Our experiments demonstrate the\nsuperiority of our method in terms of effectiveness, generalization ability,\nmodel efficiency, and sample efficiency. Our MI-HGNN improved the performance\nof a state-of-the-art model that leverages robot morphological symmetry by 8.4%\nwith only 0.21% of its parameters. Although MI-HGNN is applied to contact\nperception problems for legged robots in this work, it can be seamlessly\napplied to other types of multi-body dynamical systems and has the potential to\nimprove other robot learning frameworks. Our code is made publicly available at\nhttps://github.com/lunarlab-gatech/Morphology-Informed-HGNN.","PeriodicalId":501031,"journal":{"name":"arXiv - CS - Robotics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"MI-HGNN: Morphology-Informed Heterogeneous Graph Neural Network for Legged Robot Contact Perception\",\"authors\":\"Daniel Butterfield, Sandilya Sai Garimella, Nai-Jen Cheng, Lu Gan\",\"doi\":\"arxiv-2409.11146\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We present a Morphology-Informed Heterogeneous Graph Neural Network (MI-HGNN)\\nfor learning-based contact perception. The architecture and connectivity of the\\nMI-HGNN are constructed from the robot morphology, in which nodes and edges are\\nrobot joints and links, respectively. By incorporating the morphology-informed\\nconstraints into a neural network, we improve a learning-based approach using\\nmodel-based knowledge. We apply the proposed MI-HGNN to two contact perception\\nproblems, and conduct extensive experiments using both real-world and simulated\\ndata collected using two quadruped robots. Our experiments demonstrate the\\nsuperiority of our method in terms of effectiveness, generalization ability,\\nmodel efficiency, and sample efficiency. Our MI-HGNN improved the performance\\nof a state-of-the-art model that leverages robot morphological symmetry by 8.4%\\nwith only 0.21% of its parameters. Although MI-HGNN is applied to contact\\nperception problems for legged robots in this work, it can be seamlessly\\napplied to other types of multi-body dynamical systems and has the potential to\\nimprove other robot learning frameworks. Our code is made publicly available at\\nhttps://github.com/lunarlab-gatech/Morphology-Informed-HGNN.\",\"PeriodicalId\":501031,\"journal\":{\"name\":\"arXiv - CS - Robotics\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-17\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Robotics\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.11146\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Robotics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.11146","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
We present a Morphology-Informed Heterogeneous Graph Neural Network (MI-HGNN)
for learning-based contact perception. The architecture and connectivity of the
MI-HGNN are constructed from the robot morphology, in which nodes and edges are
robot joints and links, respectively. By incorporating the morphology-informed
constraints into a neural network, we improve a learning-based approach using
model-based knowledge. We apply the proposed MI-HGNN to two contact perception
problems, and conduct extensive experiments using both real-world and simulated
data collected using two quadruped robots. Our experiments demonstrate the
superiority of our method in terms of effectiveness, generalization ability,
model efficiency, and sample efficiency. Our MI-HGNN improved the performance
of a state-of-the-art model that leverages robot morphological symmetry by 8.4%
with only 0.21% of its parameters. Although MI-HGNN is applied to contact
perception problems for legged robots in this work, it can be seamlessly
applied to other types of multi-body dynamical systems and has the potential to
improve other robot learning frameworks. Our code is made publicly available at
https://github.com/lunarlab-gatech/Morphology-Informed-HGNN.