C-CROC: Continuous and Convex Resolution of Centroidal Dynamic Trajectories for Legged Robots in Multicontact Scenarios
- Submitting institution
-
University of Edinburgh
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 163130053
- Type
- D - Journal article
- DOI
-
10.1109/TRO.2020.2964787
- Title of journal
- IEEE Transactions on Robotics
- Article number
- -
- First page
- 676
- Volume
- 36
- Issue
- 3
- ISSN
- 1552-3098
- Open access status
- Exception within 3 months of publication
- Month of publication
- January
- Year of publication
- 2020
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
D - Language, Interaction and Robotics
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- By studying the feasibility of a problem rather than solving it, we define an abstract space where its resolution becomes trivial. We transform a hard, non-linear problem into a low dimensional convex one. The approach is formally proven to be a conservative reformulation of the Newton-Euler equations. The method can be called thousands of times per second, making it suitable for rejection sampling approaches. This allowed researchers from ETH Zurich to integrate it within a reinforcement learning framework to learn a controller able to generate locomotion over complex terrains for the Anymal quadruped (Tsounis et. al ICRA 20).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -