Matches in SemOpenAlex for { <https://semopenalex.org/work/W3090443808> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3090443808 abstract "Imitation learning is a popular approach for training effective visual navigation policies. However, collecting expert demonstrations for legged robots is challenging as these robots can be hard to control, move slowly, and cannot operate continuously for long periods of time. In this work, we propose a zero-shot imitation learning framework for training a goal-driven visual navigation policy on a legged robot from human demonstrations (third-person perspective), allowing for high-quality navigation and cost-effective data collection. However, imitation learning from third-person demonstrations raises unique challenges. First, these demonstrations are captured from different camera perspectives, which we address via a feature disentanglement network (FDN) that extracts perspective-invariant state features. Second, as transition dynamics vary between systems, we reconstruct missing action labels by either building an inverse model of the robot’s dynamics in the feature space and applying it to the human demonstrations or developing a Graphic User Interface (GUI) to label human demonstrations. To train a navigation policy we use a model-based imitation learning approach with FDN and action-labeled human demonstrations. We show that our framework can learn an effective policy for a legged robot, Laikago, from human demonstrations in both simulated and real-world environments. Our approach is zero-shot as the robot never navigates the same paths during training as those at testing time. We justify our framework by performing a comparative study." @default.
- W3090443808 created "2020-10-08" @default.
- W3090443808 creator A5002971435 @default.
- W3090443808 creator A5018507768 @default.
- W3090443808 creator A5064581452 @default.
- W3090443808 creator A5065822895 @default.
- W3090443808 creator A5084870242 @default.
- W3090443808 creator A5086928709 @default.
- W3090443808 date "2020-05-01" @default.
- W3090443808 modified "2023-09-26" @default.
- W3090443808 title "Zero-shot Imitation Learning from Demonstrations for Legged Robot Visual Navigation" @default.
- W3090443808 cites W1976980249 @default.
- W3090443808 cites W1978338723 @default.
- W3090443808 cites W1997645979 @default.
- W3090443808 cites W2029143333 @default.
- W3090443808 cites W2064451896 @default.
- W3090443808 cites W2161427949 @default.
- W3090443808 cites W2169528473 @default.
- W3090443808 cites W2281096776 @default.
- W3090443808 cites W2593841437 @default.
- W3090443808 cites W2769112066 @default.
- W3090443808 cites W2798434327 @default.
- W3090443808 cites W2806163172 @default.
- W3090443808 cites W2912063360 @default.
- W3090443808 cites W2962789679 @default.
- W3090443808 cites W2962887844 @default.
- W3090443808 cites W2962894046 @default.
- W3090443808 cites W2963447367 @default.
- W3090443808 cites W2964021598 @default.
- W3090443808 cites W2964935470 @default.
- W3090443808 cites W2967059815 @default.
- W3090443808 cites W3159890710 @default.
- W3090443808 cites W4249736682 @default.
- W3090443808 doi "https://doi.org/10.1109/icra40945.2020.9196602" @default.
- W3090443808 hasPublicationYear "2020" @default.
- W3090443808 type Work @default.
- W3090443808 sameAs 3090443808 @default.
- W3090443808 citedByCount "11" @default.
- W3090443808 countsByYear W30904438082020 @default.
- W3090443808 countsByYear W30904438082021 @default.
- W3090443808 countsByYear W30904438082022 @default.
- W3090443808 countsByYear W30904438082023 @default.
- W3090443808 crossrefType "proceedings-article" @default.
- W3090443808 hasAuthorship W3090443808A5002971435 @default.
- W3090443808 hasAuthorship W3090443808A5018507768 @default.
- W3090443808 hasAuthorship W3090443808A5064581452 @default.
- W3090443808 hasAuthorship W3090443808A5065822895 @default.
- W3090443808 hasAuthorship W3090443808A5084870242 @default.
- W3090443808 hasAuthorship W3090443808A5086928709 @default.
- W3090443808 hasBestOaLocation W30904438082 @default.
- W3090443808 hasConcept C107457646 @default.
- W3090443808 hasConcept C126388530 @default.
- W3090443808 hasConcept C138885662 @default.
- W3090443808 hasConcept C154945302 @default.
- W3090443808 hasConcept C15744967 @default.
- W3090443808 hasConcept C178790620 @default.
- W3090443808 hasConcept C185592680 @default.
- W3090443808 hasConcept C2778344882 @default.
- W3090443808 hasConcept C2780813799 @default.
- W3090443808 hasConcept C31972630 @default.
- W3090443808 hasConcept C41008148 @default.
- W3090443808 hasConcept C41895202 @default.
- W3090443808 hasConcept C77805123 @default.
- W3090443808 hasConcept C90509273 @default.
- W3090443808 hasConceptScore W3090443808C107457646 @default.
- W3090443808 hasConceptScore W3090443808C126388530 @default.
- W3090443808 hasConceptScore W3090443808C138885662 @default.
- W3090443808 hasConceptScore W3090443808C154945302 @default.
- W3090443808 hasConceptScore W3090443808C15744967 @default.
- W3090443808 hasConceptScore W3090443808C178790620 @default.
- W3090443808 hasConceptScore W3090443808C185592680 @default.
- W3090443808 hasConceptScore W3090443808C2778344882 @default.
- W3090443808 hasConceptScore W3090443808C2780813799 @default.
- W3090443808 hasConceptScore W3090443808C31972630 @default.
- W3090443808 hasConceptScore W3090443808C41008148 @default.
- W3090443808 hasConceptScore W3090443808C41895202 @default.
- W3090443808 hasConceptScore W3090443808C77805123 @default.
- W3090443808 hasConceptScore W3090443808C90509273 @default.
- W3090443808 hasLocation W30904438081 @default.
- W3090443808 hasLocation W30904438082 @default.
- W3090443808 hasOpenAccess W3090443808 @default.
- W3090443808 hasPrimaryLocation W30904438081 @default.
- W3090443808 hasRelatedWork W1592626709 @default.
- W3090443808 hasRelatedWork W1992428544 @default.
- W3090443808 hasRelatedWork W2056180080 @default.
- W3090443808 hasRelatedWork W2061090284 @default.
- W3090443808 hasRelatedWork W2096980597 @default.
- W3090443808 hasRelatedWork W2125579716 @default.
- W3090443808 hasRelatedWork W2149328105 @default.
- W3090443808 hasRelatedWork W2387675639 @default.
- W3090443808 hasRelatedWork W2397188463 @default.
- W3090443808 hasRelatedWork W4236636304 @default.
- W3090443808 isParatext "false" @default.
- W3090443808 isRetracted "false" @default.
- W3090443808 magId "3090443808" @default.
- W3090443808 workType "article" @default.