Neural dynamics of robust legged robots
Legged robot control has improved in recent years with the rise of deep reinforcement learning, however, much of the underlying neural mechanisms remain difficult to interpret. Our aim is to leverage bio-inspired methods from computational neuroscience to better understand the neural activity of rob...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-04-01
|
| Series: | Frontiers in Robotics and AI |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frobt.2024.1324404/full |
| _version_ | 1797202018408857600 |
|---|---|
| author | Eugene R. Rush Christoffer Heckman Kaushik Jayaram J. Sean Humbert |
| author_facet | Eugene R. Rush Christoffer Heckman Kaushik Jayaram J. Sean Humbert |
| author_sort | Eugene R. Rush |
| collection | DOAJ |
| description | Legged robot control has improved in recent years with the rise of deep reinforcement learning, however, much of the underlying neural mechanisms remain difficult to interpret. Our aim is to leverage bio-inspired methods from computational neuroscience to better understand the neural activity of robust robot locomotion controllers. Similar to past work, we observe that terrain-based curriculum learning improves agent stability. We study the biomechanical responses and neural activity within our neural network controller by simultaneously pairing physical disturbances with targeted neural ablations. We identify an agile hip reflex that enables the robot to regain its balance and recover from lateral perturbations. Model gradients are employed to quantify the relative degree that various sensory feedback channels drive this reflexive behavior. We also find recurrent dynamics are implicated in robust behavior, and utilize sampling-based ablation methods to identify these key neurons. Our framework combines model-based and sampling-based methods for drawing causal relationships between neural network activity and robust embodied robot behavior. |
| first_indexed | 2024-04-24T07:56:46Z |
| format | Article |
| id | doaj.art-17e4d3bd91d945f9b061c03dcaa85c3c |
| institution | Directory Open Access Journal |
| issn | 2296-9144 |
| language | English |
| last_indexed | 2024-04-24T07:56:46Z |
| publishDate | 2024-04-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Robotics and AI |
| spelling | doaj.art-17e4d3bd91d945f9b061c03dcaa85c3c2024-04-18T04:59:03ZengFrontiers Media S.A.Frontiers in Robotics and AI2296-91442024-04-011110.3389/frobt.2024.13244041324404Neural dynamics of robust legged robotsEugene R. Rush0Christoffer Heckman1Kaushik Jayaram2J. Sean Humbert3Department of Mechanical Engineering, University of Colorado Boulder, Boulder, CO, United StatesDepartment of Computer Science, University of Colorado Boulder, Boulder, CO, United StatesDepartment of Mechanical Engineering, University of Colorado Boulder, Boulder, CO, United StatesDepartment of Mechanical Engineering, University of Colorado Boulder, Boulder, CO, United StatesLegged robot control has improved in recent years with the rise of deep reinforcement learning, however, much of the underlying neural mechanisms remain difficult to interpret. Our aim is to leverage bio-inspired methods from computational neuroscience to better understand the neural activity of robust robot locomotion controllers. Similar to past work, we observe that terrain-based curriculum learning improves agent stability. We study the biomechanical responses and neural activity within our neural network controller by simultaneously pairing physical disturbances with targeted neural ablations. We identify an agile hip reflex that enables the robot to regain its balance and recover from lateral perturbations. Model gradients are employed to quantify the relative degree that various sensory feedback channels drive this reflexive behavior. We also find recurrent dynamics are implicated in robust behavior, and utilize sampling-based ablation methods to identify these key neurons. Our framework combines model-based and sampling-based methods for drawing causal relationships between neural network activity and robust embodied robot behavior.https://www.frontiersin.org/articles/10.3389/frobt.2024.1324404/fullroboticslocomotionrobustnessneurosciencereinforcement learning |
| spellingShingle | Eugene R. Rush Christoffer Heckman Kaushik Jayaram J. Sean Humbert Neural dynamics of robust legged robots Frontiers in Robotics and AI robotics locomotion robustness neuroscience reinforcement learning |
| title | Neural dynamics of robust legged robots |
| title_full | Neural dynamics of robust legged robots |
| title_fullStr | Neural dynamics of robust legged robots |
| title_full_unstemmed | Neural dynamics of robust legged robots |
| title_short | Neural dynamics of robust legged robots |
| title_sort | neural dynamics of robust legged robots |
| topic | robotics locomotion robustness neuroscience reinforcement learning |
| url | https://www.frontiersin.org/articles/10.3389/frobt.2024.1324404/full |
| work_keys_str_mv | AT eugenerrush neuraldynamicsofrobustleggedrobots AT christofferheckman neuraldynamicsofrobustleggedrobots AT kaushikjayaram neuraldynamicsofrobustleggedrobots AT jseanhumbert neuraldynamicsofrobustleggedrobots |