Stability and Safety Learning Methods for Legged Robots
Learning-based control systems have shown impressive empirical performance on challenging problems in all aspects of robot control and, in particular, in walking robots such as bipeds and quadrupeds. Unfortunately, these methods have a major critical drawback: a reduced lack of guarantees for safety...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-01-01
|
| Series: | Robotics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2218-6581/13/1/17 |
| _version_ | 1827371169945223168 |
|---|---|
| author | Paolo Arena Alessia Li Noce Luca Patanè |
| author_facet | Paolo Arena Alessia Li Noce Luca Patanè |
| author_sort | Paolo Arena |
| collection | DOAJ |
| description | Learning-based control systems have shown impressive empirical performance on challenging problems in all aspects of robot control and, in particular, in walking robots such as bipeds and quadrupeds. Unfortunately, these methods have a major critical drawback: a reduced lack of guarantees for safety and stability. In recent years, new techniques have emerged to obtain these guarantees thanks to data-driven methods that allow learning certificates together with control strategies. These techniques allow the user to verify the safety of a trained controller while providing supervision during training so that safety and stability requirements can directly influence the training process. This survey presents a comprehensive and up-to-date study of the evolving field of stability certification of neural controllers taking into account such certificates as Lyapunov functions and barrier functions. Although specific attention is paid to legged robots, several promising strategies for learning certificates, not yet applied to walking machines, are also reviewed. |
| first_indexed | 2024-03-08T10:35:22Z |
| format | Article |
| id | doaj.art-5a1515e11599472eafaabec3cf1b6a92 |
| institution | Directory Open Access Journal |
| issn | 2218-6581 |
| language | English |
| last_indexed | 2024-03-08T10:35:22Z |
| publishDate | 2024-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Robotics |
| spelling | doaj.art-5a1515e11599472eafaabec3cf1b6a922024-01-26T18:21:49ZengMDPI AGRobotics2218-65812024-01-011311710.3390/robotics13010017Stability and Safety Learning Methods for Legged RobotsPaolo Arena0Alessia Li Noce1Luca Patanè2Department of Electrical, Electronic, and Computer Engineering (DIEEI), University of Catania, 95125 Catania, ItalyDepartment of Electrical, Electronic, and Computer Engineering (DIEEI), University of Catania, 95125 Catania, ItalyDepartment of Engineering, University of Messina, 98122 Messina, ItalyLearning-based control systems have shown impressive empirical performance on challenging problems in all aspects of robot control and, in particular, in walking robots such as bipeds and quadrupeds. Unfortunately, these methods have a major critical drawback: a reduced lack of guarantees for safety and stability. In recent years, new techniques have emerged to obtain these guarantees thanks to data-driven methods that allow learning certificates together with control strategies. These techniques allow the user to verify the safety of a trained controller while providing supervision during training so that safety and stability requirements can directly influence the training process. This survey presents a comprehensive and up-to-date study of the evolving field of stability certification of neural controllers taking into account such certificates as Lyapunov functions and barrier functions. Although specific attention is paid to legged robots, several promising strategies for learning certificates, not yet applied to walking machines, are also reviewed.https://www.mdpi.com/2218-6581/13/1/17legged robotsstabilityneural networksreinforcement learninglocomotion control |
| spellingShingle | Paolo Arena Alessia Li Noce Luca Patanè Stability and Safety Learning Methods for Legged Robots Robotics legged robots stability neural networks reinforcement learning locomotion control |
| title | Stability and Safety Learning Methods for Legged Robots |
| title_full | Stability and Safety Learning Methods for Legged Robots |
| title_fullStr | Stability and Safety Learning Methods for Legged Robots |
| title_full_unstemmed | Stability and Safety Learning Methods for Legged Robots |
| title_short | Stability and Safety Learning Methods for Legged Robots |
| title_sort | stability and safety learning methods for legged robots |
| topic | legged robots stability neural networks reinforcement learning locomotion control |
| url | https://www.mdpi.com/2218-6581/13/1/17 |
| work_keys_str_mv | AT paoloarena stabilityandsafetylearningmethodsforleggedrobots AT alessialinoce stabilityandsafetylearningmethodsforleggedrobots AT lucapatane stabilityandsafetylearningmethodsforleggedrobots |