Feng Gao, Chengjia Lei, Xingguo Long, Jin Wang, Peiheng Song
{"title":"Design and Development of an Intelligent Pet-Type Quadruped Robot","authors":"Feng Gao, Chengjia Lei, Xingguo Long, Jin Wang, Peiheng Song","doi":"10.1109/MIPR51284.2021.00068","DOIUrl":null,"url":null,"abstract":"Inspired by the assistance that artificial intelligence offers to artistic creation, we apply AI technology to create the Open Monster C class number 01 (OM-C01), a quadruped robot dog as lifelike as an artwork. OM-C01 adopts a 2-DoF five-bar parallel mechanism to realize the thigh and shank bionic structure. We combine the visual learning system based on few-shot learning and incremental learning with GPT-2 pre-training language model to endow OM-C01 the same learning ability as a pet. OM-C01 can make decisions based on the facial expression as well as its emotional state, and shape a unique personality by updating the Q-table. Meanwhile, we implement a digital twin simulation environment for OM-C01 based on .NET WPF, which is convenient for designing various actions.","PeriodicalId":139543,"journal":{"name":"2021 IEEE 4th International Conference on Multimedia Information Processing and Retrieval (MIPR)","volume":"334 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE 4th International Conference on Multimedia Information Processing and Retrieval (MIPR)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/MIPR51284.2021.00068","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
Inspired by the assistance that artificial intelligence offers to artistic creation, we apply AI technology to create the Open Monster C class number 01 (OM-C01), a quadruped robot dog as lifelike as an artwork. OM-C01 adopts a 2-DoF five-bar parallel mechanism to realize the thigh and shank bionic structure. We combine the visual learning system based on few-shot learning and incremental learning with GPT-2 pre-training language model to endow OM-C01 the same learning ability as a pet. OM-C01 can make decisions based on the facial expression as well as its emotional state, and shape a unique personality by updating the Q-table. Meanwhile, we implement a digital twin simulation environment for OM-C01 based on .NET WPF, which is convenient for designing various actions.