Automated tuning of nonlinear model predictive controller by reinforcement learning

One of the major challenges of model predictive control (MPC) for robotic applications is the non-trivial weight tuning process while crafting the objective function. This process is often executed using the trial-and-error method by the user. Consequently, the optimality of the weights and the time...

Full description

Saved in:

Bibliographic Details
Main Authors:	Mehndiratta, Mohit, Camci, Efe, Kayacan, Erdal
Other Authors:	School of Mechanical and Aerospace Engineering
Format:	Conference or Workshop Item
Language:	English
Published:	2020
Subjects:	Engineering::Mechanical engineering Tuning Rotors
Online Access:	https://hdl.handle.net/10356/143042
Tags:	Add Tag No Tags, Be the first to tag this record!
Institution:	Nanyang Technological University
Language:	English

id	sg-ntu-dr.10356-143042
record_format	dspace
spelling	sg-ntu-dr.10356-1430422023-03-04T17:07:08Z Automated tuning of nonlinear model predictive controller by reinforcement learning Mehndiratta, Mohit Camci, Efe Kayacan, Erdal School of Mechanical and Aerospace Engineering 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Engineering::Mechanical engineering Tuning Rotors One of the major challenges of model predictive control (MPC) for robotic applications is the non-trivial weight tuning process while crafting the objective function. This process is often executed using the trial-and-error method by the user. Consequently, the optimality of the weights and the time required for the process become highly dependent on the skill set and experience of the user. In this study, we present a generic and user-independent framework which automates the tuning process by reinforcement learning. The proposed method shows competency in tuning a nonlinear MPC (NMPC) which is employed for trajectory tracking control of aerial robots. It explores the desirable weights within less than an hour in iterative Gazebo simulations running on a standard desktop computer. The real world experiments illustrate that the NMPC weights explored by the proposed method result in a satisfactory trajectory tracking performance. Ministry of Education (MOE) Accepted version This work was financially supported by the Singapore Ministry of Education (RG185/17) and Aarhus University, Department of Engineering (28173). 2020-07-22T08:36:13Z 2020-07-22T08:36:13Z 2019 Conference Paper Mehndiratta, M., Camci, E., & Kayacan, E. (2018). Automated tuning of nonlinear model predictive controller by reinforcement learning. Proceedings of 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 3016-3021. doi:10.1109/iros.2018.8594350 978-1-5386-8095-7 https://hdl.handle.net/10356/143042 10.1109/iros.2018.8594350 2-s2.0-85062939914 3016 3021 en © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/iros.2018.8594350 application/pdf
institution	Nanyang Technological University
building	NTU Library
continent	Asia
country	Singapore Singapore
content_provider	NTU Library
collection	DR-NTU
language	English
topic	Engineering::Mechanical engineering Tuning Rotors
spellingShingle	Engineering::Mechanical engineering Tuning Rotors Mehndiratta, Mohit Camci, Efe Kayacan, Erdal Automated tuning of nonlinear model predictive controller by reinforcement learning
description	One of the major challenges of model predictive control (MPC) for robotic applications is the non-trivial weight tuning process while crafting the objective function. This process is often executed using the trial-and-error method by the user. Consequently, the optimality of the weights and the time required for the process become highly dependent on the skill set and experience of the user. In this study, we present a generic and user-independent framework which automates the tuning process by reinforcement learning. The proposed method shows competency in tuning a nonlinear MPC (NMPC) which is employed for trajectory tracking control of aerial robots. It explores the desirable weights within less than an hour in iterative Gazebo simulations running on a standard desktop computer. The real world experiments illustrate that the NMPC weights explored by the proposed method result in a satisfactory trajectory tracking performance.
author2	School of Mechanical and Aerospace Engineering
author_facet	School of Mechanical and Aerospace Engineering Mehndiratta, Mohit Camci, Efe Kayacan, Erdal
format	Conference or Workshop Item
author	Mehndiratta, Mohit Camci, Efe Kayacan, Erdal
author_sort	Mehndiratta, Mohit
title	Automated tuning of nonlinear model predictive controller by reinforcement learning
title_short	Automated tuning of nonlinear model predictive controller by reinforcement learning
title_full	Automated tuning of nonlinear model predictive controller by reinforcement learning
title_fullStr	Automated tuning of nonlinear model predictive controller by reinforcement learning
title_full_unstemmed	Automated tuning of nonlinear model predictive controller by reinforcement learning
title_sort	automated tuning of nonlinear model predictive controller by reinforcement learning
publishDate	2020
url	https://hdl.handle.net/10356/143042
_version_	1759853536406929408

Automated tuning of nonlinear model predictive controller by reinforcement learning

Similar Items