Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386302603> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W4386302603 abstract "Summarizing knowledge from animals and human beings inspires robotic innovations. In this work, we propose a framework for driving legged robots act like real animals with lifelike agility and strategy in complex environments. Inspired by large pre-trained models witnessed with impressive performance in language and image understanding, we introduce the power of advanced deep generative models to produce motor control signals stimulating legged robots to act like real animals. Unlike conventional controllers and end-to-end RL methods that are task-specific, we propose to pre-train generative models over animal motion datasets to preserve expressive knowledge of animal behavior. The pre-trained model holds sufficient primitive-level knowledge yet is environment-agnostic. It is then reused for a successive stage of learning to align with the environments by traversing a number of challenging obstacles that are rarely considered in previous approaches, including creeping through narrow spaces, jumping over hurdles, freerunning over scattered blocks, etc. Finally, a task-specific controller is trained to solve complex downstream tasks by reusing the knowledge from previous stages. Enriching the knowledge regarding each stage does not affect the usage of other levels of knowledge. This flexible framework offers the possibility of continual knowledge accumulation at different levels. We successfully apply the trained multi-level controllers to the MAX robot, a quadrupedal robot developed in-house, to mimic animals, traverse complex obstacles, and play in a designed challenging multi-agent Chase Tag Game, where lifelike agility and strategy emerge on the robots. The present research pushes the frontier of robot control with new insights on reusing multi-level pre-trained knowledge and solving highly complex downstream tasks in the real world." @default.
- W4386302603 created "2023-08-31" @default.
- W4386302603 creator A5020958723 @default.
- W4386302603 creator A5023854164 @default.
- W4386302603 creator A5029261791 @default.
- W4386302603 creator A5030364414 @default.
- W4386302603 creator A5037373363 @default.
- W4386302603 creator A5038823196 @default.
- W4386302603 creator A5043755970 @default.
- W4386302603 creator A5047835985 @default.
- W4386302603 creator A5058849717 @default.
- W4386302603 creator A5062755510 @default.
- W4386302603 creator A5071563180 @default.
- W4386302603 creator A5075596275 @default.
- W4386302603 creator A5078304976 @default.
- W4386302603 creator A5079592609 @default.
- W4386302603 creator A5081520768 @default.
- W4386302603 creator A5083545249 @default.
- W4386302603 creator A5085352453 @default.
- W4386302603 creator A5087050304 @default.
- W4386302603 creator A5090560809 @default.
- W4386302603 date "2023-08-29" @default.
- W4386302603 modified "2023-09-27" @default.
- W4386302603 title "Lifelike Agility and Play on Quadrupedal Robots using Reinforcement Learning and Generative Pre-trained Models" @default.
- W4386302603 doi "https://doi.org/10.48550/arxiv.2308.15143" @default.
- W4386302603 hasPublicationYear "2023" @default.
- W4386302603 type Work @default.
- W4386302603 citedByCount "0" @default.
- W4386302603 crossrefType "posted-content" @default.
- W4386302603 hasAuthorship W4386302603A5020958723 @default.
- W4386302603 hasAuthorship W4386302603A5023854164 @default.
- W4386302603 hasAuthorship W4386302603A5029261791 @default.
- W4386302603 hasAuthorship W4386302603A5030364414 @default.
- W4386302603 hasAuthorship W4386302603A5037373363 @default.
- W4386302603 hasAuthorship W4386302603A5038823196 @default.
- W4386302603 hasAuthorship W4386302603A5043755970 @default.
- W4386302603 hasAuthorship W4386302603A5047835985 @default.
- W4386302603 hasAuthorship W4386302603A5058849717 @default.
- W4386302603 hasAuthorship W4386302603A5062755510 @default.
- W4386302603 hasAuthorship W4386302603A5071563180 @default.
- W4386302603 hasAuthorship W4386302603A5075596275 @default.
- W4386302603 hasAuthorship W4386302603A5078304976 @default.
- W4386302603 hasAuthorship W4386302603A5079592609 @default.
- W4386302603 hasAuthorship W4386302603A5081520768 @default.
- W4386302603 hasAuthorship W4386302603A5083545249 @default.
- W4386302603 hasAuthorship W4386302603A5085352453 @default.
- W4386302603 hasAuthorship W4386302603A5087050304 @default.
- W4386302603 hasAuthorship W4386302603A5090560809 @default.
- W4386302603 hasBestOaLocation W43863026031 @default.
- W4386302603 hasConcept C107457646 @default.
- W4386302603 hasConcept C127413603 @default.
- W4386302603 hasConcept C13280743 @default.
- W4386302603 hasConcept C154945302 @default.
- W4386302603 hasConcept C167966045 @default.
- W4386302603 hasConcept C176809094 @default.
- W4386302603 hasConcept C201995342 @default.
- W4386302603 hasConcept C205649164 @default.
- W4386302603 hasConcept C2780451532 @default.
- W4386302603 hasConcept C39890363 @default.
- W4386302603 hasConcept C41008148 @default.
- W4386302603 hasConcept C90509273 @default.
- W4386302603 hasConcept C97541855 @default.
- W4386302603 hasConceptScore W4386302603C107457646 @default.
- W4386302603 hasConceptScore W4386302603C127413603 @default.
- W4386302603 hasConceptScore W4386302603C13280743 @default.
- W4386302603 hasConceptScore W4386302603C154945302 @default.
- W4386302603 hasConceptScore W4386302603C167966045 @default.
- W4386302603 hasConceptScore W4386302603C176809094 @default.
- W4386302603 hasConceptScore W4386302603C201995342 @default.
- W4386302603 hasConceptScore W4386302603C205649164 @default.
- W4386302603 hasConceptScore W4386302603C2780451532 @default.
- W4386302603 hasConceptScore W4386302603C39890363 @default.
- W4386302603 hasConceptScore W4386302603C41008148 @default.
- W4386302603 hasConceptScore W4386302603C90509273 @default.
- W4386302603 hasConceptScore W4386302603C97541855 @default.
- W4386302603 hasLocation W43863026031 @default.
- W4386302603 hasOpenAccess W4386302603 @default.
- W4386302603 hasPrimaryLocation W43863026031 @default.
- W4386302603 hasRelatedWork W1763389228 @default.
- W4386302603 hasRelatedWork W2007682987 @default.
- W4386302603 hasRelatedWork W2079554071 @default.
- W4386302603 hasRelatedWork W2166791242 @default.
- W4386302603 hasRelatedWork W2323122434 @default.
- W4386302603 hasRelatedWork W2343019076 @default.
- W4386302603 hasRelatedWork W3038859464 @default.
- W4386302603 hasRelatedWork W3186584605 @default.
- W4386302603 hasRelatedWork W4229726131 @default.
- W4386302603 hasRelatedWork W4385572368 @default.
- W4386302603 isParatext "false" @default.
- W4386302603 isRetracted "false" @default.
- W4386302603 workType "article" @default.