Interpretable Active Inference Gait Control Learning

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F25%3A00388320" target="_blank" >RIV/68407700:21230/25:00388320 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1109/ICRA55743.2025.11128724" target="_blank" >https://doi.org/10.1109/ICRA55743.2025.11128724</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICRA55743.2025.11128724" target="_blank" >10.1109/ICRA55743.2025.11128724</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Interpretable Active Inference Gait Control Learning

Popis výsledku v původním jazyce

Sustaining the gait locomotion in an adversarial environment requires the robot to react to novel experiences adaptively. In Free Energy Principle (FEP), the behavioral reaction is driven by the discrepancy between observation and prediction. Although, for legged robot gait locomotion, the prediction of gait dynamics is challenging as the consequences non-linearly depend on the activity history, the animal gait is robust, adapting to severe motion disruptions seemingly instantly. In biomimetic robotics, the Central Pattern Generator (CPG) relaxes the general dynamics of body-environment interaction to the stable and repetitive dynamics of gait. Based on these observations, we propose self-learning of the gait dynamics model and FEP framework that infers state estimation and gait control. The proposed method is experimentally evaluated on a real hexapod walking robot with 18 controllable degrees of freedom. The robot learns the gait dynamics model indoors and then deploys it in outdoor navigation under various adversarial scenarios. Results show that the developed interpretable gait controller exhibits complex and real-time adaptive behavior when it encounters unknown situations.

Název v anglickém jazyce

Interpretable Active Inference Gait Control Learning

Popis výsledku anglicky

Sustaining the gait locomotion in an adversarial environment requires the robot to react to novel experiences adaptively. In Free Energy Principle (FEP), the behavioral reaction is driven by the discrepancy between observation and prediction. Although, for legged robot gait locomotion, the prediction of gait dynamics is challenging as the consequences non-linearly depend on the activity history, the animal gait is robust, adapting to severe motion disruptions seemingly instantly. In biomimetic robotics, the Central Pattern Generator (CPG) relaxes the general dynamics of body-environment interaction to the stable and repetitive dynamics of gait. Based on these observations, we propose self-learning of the gait dynamics model and FEP framework that infers state estimation and gait control. The proposed method is experimentally evaluated on a real hexapod walking robot with 18 controllable degrees of freedom. The robot learns the gait dynamics model indoors and then deploys it in outdoor navigation under various adversarial scenarios. Results show that the developed interpretable gait controller exhibits complex and real-time adaptive behavior when it encounters unknown situations.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20204 - Robotics and automatic control

Projekt

Výsledek vznikl pri realizaci vícero projektů. Více informací v záložce Projekty.

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

2025 IEEE International Conference on Robotics and Automation (ICRA)

ISBN

979-8-3315-4139-2

ISSN

1050-4729

e-ISSN

1050-4729

Počet stran výsledku

7

Strana od-do

9630-9636

Název nakladatele

IEEE Industrial Electronic Society

Místo vydání

Vienna

Místo konání akce

Atlanta

Datum konání akce

19. 5. 2025

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001614845800364