Collaborative source-seeking control

Ruggero Fabbiano 1, 2
2 NECS - Networked Controlled Systems
Inria Grenoble - Rhône-Alpes, GIPSA-DA - Département Automatique
Abstract : The dissertation faces the problem of source localisation, a topic which has been extensively studied in recent literature due to its large number of applications. In particular, it focuses on steering multiple sensors, able to take point-wise measurements of the emitted quantity, towards the source without making use of any position information, which happens to be unavailable in many practical cases (for example, underwater or underground exploration). By making some assumptions on the diffusion process, we develop a model which allows us to use some mathematical tools (the Poisson integral and its derivatives) for a simple approximation of the gradient of the function describing the diffusion process, whose source represents its maximum, making it possible to perform a gradient ascent to find the source location. The contributions are threefold: first, we use such tools to solve a 2-dimensional centralised source-seeking problem, where a single vehicle, equipped with multiple sensors and without position information, is moving in a planar environment where a source is supposed to emit. Then, we extend it to a 3-dimensional framework, considering a flying vehicle equipped with sensors moving in the space; for this more general case, in addition to simulation validation, we provide a theoretical study of the convergence properties of the proposed control law. Finally, we tackle the distributed source-localisation problem, considering several autonomous moving sensors (in two dimensions); in addition to the problem of implementing the source-localisation algorithm in a distributed manner, in this latter case we have also to guarantee a suitable formation control, to ensure the correctness of the gradient estimation and hence reach the source.
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-01217117
Contributor : Abes Star <>
Submitted on : Monday, October 19, 2015 - 9:58:06 AM
Last modification on : Friday, September 27, 2019 - 10:43:40 PM
Long-term archiving on : Wednesday, January 20, 2016 - 11:21:41 AM

File

FABBIANO_2015_archivage.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01217117, version 1

Citation

Ruggero Fabbiano. Collaborative source-seeking control. Signal and Image processing. Université Grenoble Alpes, 2015. English. ⟨NNT : 2015GREAT038⟩. ⟨tel-01217117⟩

Share

Metrics

Record views

1047

Files downloads

619