Sampling-Based Model Predictive Control Leveraging Parallelizable Physics Simulations

Corrado Pezzato; Chadi Salmi; Elia Trevisan; Max Spahn; Javier Alonso-Mora; Carlos Hernández Corbato

doi:10.1109/LRA.2025.3535185

Sampling-Based Model Predictive Control Leveraging Parallelizable Physics Simulations

Corrado Pezzato, Chadi Salmi, Elia Trevisan^*, Max Spahn, Javier Alonso-Mora, Carlos Hernández Corbato

^*Corresponding author for this work

Research output: Contribution to journal › Article › Scientific › peer-review

Abstract

We present a sampling-based model predictive control method that uses a generic physics simulator as the dynamical model. In particular, we propose a Model Predictive Path Integral controller (MPPI) that employs the GPU-parallelizable IsaacGym simulator to compute the forward dynamics of the robot and environment. Since the simulator implicitly defines the dynamic model, our method is readily extendable to different objects and robots, allowing one to solve complex navigation and contact-rich tasks. We demonstrate the effectiveness of this method in several simulated and real-world settings, including mobile navigation with collision avoidance, non-prehensile manipulation, and whole-body control for high-dimensional configuration spaces. This is a powerful and accessible open-source tool to solve many contact-rich motion planning tasks.

Original language	English
Pages (from-to)	2750-2757
Number of pages	8
Journal	IEEE Robotics and Automation Letters
Volume	10
Issue number	3
DOIs	https://doi.org/10.1109/LRA.2025.3535185
Publication status	Published - 2025

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

Contact Modeling
Model Predictive Path Integral Control
Optimization and Optimal Control
Whole-Body Motion Planning and Control

Access to Document

10.1109/LRA.2025.3535185

Cite this

@article{46bf05b3f1d94fb89f510188125e882c,

title = "Sampling-Based Model Predictive Control Leveraging Parallelizable Physics Simulations",

abstract = "We present a sampling-based model predictive control method that uses a generic physics simulator as the dynamical model. In particular, we propose a Model Predictive Path Integral controller (MPPI) that employs the GPU-parallelizable IsaacGym simulator to compute the forward dynamics of the robot and environment. Since the simulator implicitly defines the dynamic model, our method is readily extendable to different objects and robots, allowing one to solve complex navigation and contact-rich tasks. We demonstrate the effectiveness of this method in several simulated and real-world settings, including mobile navigation with collision avoidance, non-prehensile manipulation, and whole-body control for high-dimensional configuration spaces. This is a powerful and accessible open-source tool to solve many contact-rich motion planning tasks.",

keywords = "Contact Modeling, Model Predictive Path Integral Control, Optimization and Optimal Control, Whole-Body Motion Planning and Control",

author = "Corrado Pezzato and Chadi Salmi and Elia Trevisan and Max Spahn and Javier Alonso-Mora and Corbato, {Carlos Hern{\'a}ndez}",

note = "Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.",

year = "2025",

doi = "10.1109/LRA.2025.3535185",

language = "English",

volume = "10",

pages = "2750--2757",

journal = "IEEE Robotics and Automation Letters",

issn = "2377-3766",

publisher = "IEEE",

number = "3",

}

TY - JOUR

T1 - Sampling-Based Model Predictive Control Leveraging Parallelizable Physics Simulations

AU - Pezzato, Corrado

AU - Salmi, Chadi

AU - Trevisan, Elia

AU - Spahn, Max

AU - Alonso-Mora, Javier

AU - Corbato, Carlos Hernández

N1 - Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2025

Y1 - 2025

N2 - We present a sampling-based model predictive control method that uses a generic physics simulator as the dynamical model. In particular, we propose a Model Predictive Path Integral controller (MPPI) that employs the GPU-parallelizable IsaacGym simulator to compute the forward dynamics of the robot and environment. Since the simulator implicitly defines the dynamic model, our method is readily extendable to different objects and robots, allowing one to solve complex navigation and contact-rich tasks. We demonstrate the effectiveness of this method in several simulated and real-world settings, including mobile navigation with collision avoidance, non-prehensile manipulation, and whole-body control for high-dimensional configuration spaces. This is a powerful and accessible open-source tool to solve many contact-rich motion planning tasks.

AB - We present a sampling-based model predictive control method that uses a generic physics simulator as the dynamical model. In particular, we propose a Model Predictive Path Integral controller (MPPI) that employs the GPU-parallelizable IsaacGym simulator to compute the forward dynamics of the robot and environment. Since the simulator implicitly defines the dynamic model, our method is readily extendable to different objects and robots, allowing one to solve complex navigation and contact-rich tasks. We demonstrate the effectiveness of this method in several simulated and real-world settings, including mobile navigation with collision avoidance, non-prehensile manipulation, and whole-body control for high-dimensional configuration spaces. This is a powerful and accessible open-source tool to solve many contact-rich motion planning tasks.

KW - Contact Modeling

KW - Model Predictive Path Integral Control

KW - Optimization and Optimal Control

KW - Whole-Body Motion Planning and Control

UR - http://www.scopus.com/inward/record.url?scp=85217061884&partnerID=8YFLogxK

U2 - 10.1109/LRA.2025.3535185

DO - 10.1109/LRA.2025.3535185

M3 - Article

AN - SCOPUS:85217061884

SN - 2377-3766

VL - 10

SP - 2750

EP - 2757

JO - IEEE Robotics and Automation Letters

JF - IEEE Robotics and Automation Letters

IS - 3

ER -

Sampling-Based Model Predictive Control Leveraging Parallelizable Physics Simulations

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Embargoed Document

Fingerprint

Cite this