A Real-Time Approach for Humanoid Robot Walking including Dynamic Obstacles Avoidance - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

A Real-Time Approach for Humanoid Robot Walking including Dynamic Obstacles Avoidance

Résumé

This paper proposes a novel approach to online replan the walking trajectory of a biped humanoid robot to avoid unexpected interactions and impacts with dynamic obstacles that may compromise the balance of the humanoid robot. The proposed method adjusts the position of the contacts of a preplanned global trajectory according to the position of moving obstacles and the robot's dynamic properties. The methodology includes a graph-based footstep planner to generate a footstep sequence aware of possible changes in a dynamic environment, a Model Predictive Controller based on the Single-Rigid Body Dynamics to track the computed footsteps, and a final Whole-Body Control layer to compute proper joint torque commands. Preliminary results using the proposed approach are presented to demonstrate the effectiveness of the proposed framework in simulated scenarios with the DRACO3 humanoid bipedal platform.
Fichier principal
Vignette du fichier
humanoids_2023_submitted_version.pdf (5.27 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04228814 , version 1 (04-10-2023)
hal-04228814 , version 2 (20-10-2023)
hal-04228814 , version 3 (27-10-2023)

Identifiants

  • HAL Id : hal-04228814 , version 2

Citer

Luca Rossini, Enrico Mingo Hoffman, Seung Hyeon Bang, Luis Sentis, Nikos G Tsagarakis. A Real-Time Approach for Humanoid Robot Walking including Dynamic Obstacles Avoidance. IEEE International Conference on Humanoid Robots (HUMANOIDS) 2023, Dec 2023, Austin, United States. ⟨hal-04228814v2⟩
210 Consultations
193 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More