Gradient-based iterative learning control for decentralised collaborative tracking
Gradient-based iterative learning control for decentralised collaborative tracking
Collaborative tracking control of multi-agent systems involves two or more subsystems working together to perform a global objective, and is increasingly used within a diverse range of applications. However existing, predominately centralised, control structures are sensitive to communication delays and data drop-out leading to inaccurate tracking. Iterative learning control (ILC) has been applied to increase performance using past experience, but reliance on inverse dynamics has severely reduced robustness to model uncertainty. This paper proposes the first general decentralized iterative learning framework to address this problem, thereby enabling a wide range of existing ILC methodologies to be applied to this area. This framework is illustrated through the derivation of a decentralised gradient based ILC algorithm which ensures convergence to the required reference trajectory, while simultaneously optimising the control input energy. In addition, a novel balancing algorithm is also proposed to distribute the input energy of each agent and hence avoid sub agent overloading.
721-726
Chen, Shangcheng
ad9127e5-deb9-48be-bc8b-c8719f2c023a
Freeman, Christopher
ccdd1272-cdc7-43fb-a1bb-b1ef0bdf5815
Chu, Bing
555a86a5-0198-4242-8525-3492349d4f0f
27 November 2018
Chen, Shangcheng
ad9127e5-deb9-48be-bc8b-c8719f2c023a
Freeman, Christopher
ccdd1272-cdc7-43fb-a1bb-b1ef0bdf5815
Chu, Bing
555a86a5-0198-4242-8525-3492349d4f0f
Chen, Shangcheng, Freeman, Christopher and Chu, Bing
(2018)
Gradient-based iterative learning control for decentralised collaborative tracking.
In 2018 European Control Conference, ECC 2018.
IEEE.
.
(doi:10.23919/ECC.2018.8550177).
Record type:
Conference or Workshop Item
(Paper)
Abstract
Collaborative tracking control of multi-agent systems involves two or more subsystems working together to perform a global objective, and is increasingly used within a diverse range of applications. However existing, predominately centralised, control structures are sensitive to communication delays and data drop-out leading to inaccurate tracking. Iterative learning control (ILC) has been applied to increase performance using past experience, but reliance on inverse dynamics has severely reduced robustness to model uncertainty. This paper proposes the first general decentralized iterative learning framework to address this problem, thereby enabling a wide range of existing ILC methodologies to be applied to this area. This framework is illustrated through the derivation of a decentralised gradient based ILC algorithm which ensures convergence to the required reference trajectory, while simultaneously optimising the control input energy. In addition, a novel balancing algorithm is also proposed to distribute the input energy of each agent and hence avoid sub agent overloading.
This record has no associated files available for download.
More information
Published date: 27 November 2018
Venue - Dates:
16th European Control Conference, ECC 2018, , Limassol, Cyprus, 2018-06-12 - 2018-06-15
Identifiers
Local EPrints ID: 427609
URI: http://eprints.soton.ac.uk/id/eprint/427609
PURE UUID: c2c6a811-9b89-47f1-ae5d-1a42cbf044a6
Catalogue record
Date deposited: 24 Jan 2019 17:30
Last modified: 11 Dec 2024 02:39
Export record
Altmetrics
Contributors
Author:
Shangcheng Chen
Author:
Christopher Freeman
Author:
Bing Chu
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics