Optimization for non-periodic dynamic motions of legged systems
Autonomous legged robots will be required to handle a wide variety of tasks and environments. While a lot of research has focused on developing their abilities for periodic locomotion tasks, little effort has been invested in devising generalized strategies for dynamic, non-periodic movements. In th...
| Autors principals: | , , , |
|---|---|
| Format: | Conference item |
| Publicat: |
International Workshop on Human Friendly Robotics
2016
|
| _version_ | 1826257174653829120 |
|---|---|
| author | Radulescu, A Havoutis, I Caldwell, D Semini, C |
| author_facet | Radulescu, A Havoutis, I Caldwell, D Semini, C |
| author_sort | Radulescu, A |
| collection | OXFORD |
| description | Autonomous legged robots will be required to handle a wide variety of tasks and environments. While a lot of research has focused on developing their abilities for periodic locomotion tasks, little effort has been invested in devising generalized strategies for dynamic, non-periodic movements. In this work we employ the use of optimization and learning to overcome the use of motion design approaches, frequently by teleoperation, in such scenarios. We employ a realistic simulation of the HyQ2Max quadrupedal system for investigations on two distinctive tasks: rearing and posture recovery. The results obtained show the potential of optimization and learning approaches for motion synthesis in the context of complex tasks. |
| first_indexed | 2024-03-06T18:13:56Z |
| format | Conference item |
| id | oxford-uuid:03f6c077-968a-45f6-ab47-f5050e44b597 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T18:13:56Z |
| publishDate | 2016 |
| publisher | International Workshop on Human Friendly Robotics |
| record_format | dspace |
| spelling | oxford-uuid:03f6c077-968a-45f6-ab47-f5050e44b5972022-03-26T08:49:18ZOptimization for non-periodic dynamic motions of legged systemsConference itemhttp://purl.org/coar/resource_type/c_5794uuid:03f6c077-968a-45f6-ab47-f5050e44b597Symplectic Elements at OxfordInternational Workshop on Human Friendly Robotics2016Radulescu, AHavoutis, ICaldwell, DSemini, CAutonomous legged robots will be required to handle a wide variety of tasks and environments. While a lot of research has focused on developing their abilities for periodic locomotion tasks, little effort has been invested in devising generalized strategies for dynamic, non-periodic movements. In this work we employ the use of optimization and learning to overcome the use of motion design approaches, frequently by teleoperation, in such scenarios. We employ a realistic simulation of the HyQ2Max quadrupedal system for investigations on two distinctive tasks: rearing and posture recovery. The results obtained show the potential of optimization and learning approaches for motion synthesis in the context of complex tasks. |
| spellingShingle | Radulescu, A Havoutis, I Caldwell, D Semini, C Optimization for non-periodic dynamic motions of legged systems |
| title | Optimization for non-periodic dynamic motions of legged systems |
| title_full | Optimization for non-periodic dynamic motions of legged systems |
| title_fullStr | Optimization for non-periodic dynamic motions of legged systems |
| title_full_unstemmed | Optimization for non-periodic dynamic motions of legged systems |
| title_short | Optimization for non-periodic dynamic motions of legged systems |
| title_sort | optimization for non periodic dynamic motions of legged systems |
| work_keys_str_mv | AT radulescua optimizationfornonperiodicdynamicmotionsofleggedsystems AT havoutisi optimizationfornonperiodicdynamicmotionsofleggedsystems AT caldwelld optimizationfornonperiodicdynamicmotionsofleggedsystems AT seminic optimizationfornonperiodicdynamicmotionsofleggedsystems |