Global Corridor-Guided Adaptive Covariance MPPI for a Robotic Manipulator in Constrained Environments
IEEE Access · 2026 · 14 · 124079-124092
DOI: 10.1109/access.2026.3721129 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Munhaeng Lee und 2 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Trade, Industry & Energy (MOTIE, Korea) under the Industrial Technology Innovation Program.
Abstract (Originalsprache)
This paper proposes a global corridor-guided adaptive covariance model predictive path integral (MPPI) framework for pose control of robotic manipulators in constrained environments. The proposed method converts a global path generated by RRT-Connect into a sphere-based corridor and uses the corridor as a feasible guidance region for the tool center point (TCP) instead of directly enforcing path tracking. It then introduces a corridor-based guiding vector (CGV) inside the corridor to provide directional guidance for detouring motion and to alleviate stagnation caused by corridor-only guidance. In addition, it improves the balance between exploration and convergence behavior by dynamically adjusting the covariance based on the task progress and corridor size. Numerical simulations with a 6-DoF robotic manipulator validate the proposed framework. The results show that the proposed method improves convergence behavior and reduces oscillatory motion near the goal pose compared with fixed-covariance MPPI variants.
Förderung
- Ministry of Trade, Industry & Energy (MOTIE, Korea) under the Industrial Technology Innovation Program
Zitieren
APA
Lee, M., Kim, S., & Suh, J. (2026). Global Corridor-Guided Adaptive Covariance MPPI for a Robotic Manipulator in Constrained Environments. IEEE Access, 14, 124079-124092. https://doi.org/10.1109/access.2026.3721129
Vancouver
Lee M, Kim S, Suh J. Global Corridor-Guided Adaptive Covariance MPPI for a Robotic Manipulator in Constrained Environments. IEEE Access. 2026;14:124079-124092. doi:10.1109/access.2026.3721129
BibTeX
@article{Lee2026,
author = {Lee, Munhaeng and Kim, Sungjae and Suh, Jin-Ho},
title = {Global Corridor-Guided Adaptive Covariance MPPI for a Robotic Manipulator in Constrained Environments},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {124079--124092},
doi = {10.1109/access.2026.3721129}
}
EUVIMED