Job opportunities

3iA Postdoctoral fellowship 
24 months

Machine Learning methods for source separation
of distributed fiber optic sensing data


Optical fiber, in addition to being a means of transmitting information, is also a material that is very sensitive to environmental variations. When a laser light pulse travels through an optical fiber, it interacts with tiny impurities in the material and optical backscattering occurs. Processing this response provides estimates of the local variations in acoustic pressure along the fiber, over distances ranging from 40km up to 140km with some systems. This technique, called Distributed Acoustic Sensing (DAS), is currently experiencing growing interest in an increasing number of applications, e.g., traffic transportation monitoring, structural health monitoring, and natural hazards detection, to cite a few. As telecom fibers are ubiquitous in urban environments, DAS appears as a breakthrough concept to upgrade existing fiber optic networks to acoustic sensor arrays, and a key component for managing smart cities.
In many applications, DAS suffers from the multi-source aliasing problem and low-frequency noise, especially in noisy environments. Indeed, when multiple sources exist in the vicinity of a same sensing unit, their signatures mix and estimation of individual sources is disturbed by the other co-occurring sources. The aim of the postdoctoral research work will be to devise efficient algorithms for source separation in DAS measurements. Issues such as large data volumes that can exceed 1 To per day and per fiber, instrument noise, complex nature of the moving sources, and directionality of the DAS measurements, make the use of machine learning techniques very appealing. Different types of deep neural networks have been used recently to address the audio source separation problem including the denoising autoencoder. The post-doctoral fellow will build upon these methods for proposing deep learning methods for source separation of DAS data. Experiments will be carried out on urban, costal and underwater noisy data. The novelty of the application and the relative lack of a framework for processing DAS data should ensure fast dissemination of this work. 

Location: Lagrange Lab., Campus Valrose, University Côte d'Azur, Nice (downtown), France

To prepare your application, please send me an email
The next application step will include:
• Curriculum vitæ including the list of the scientific publications
• Motivation letter
• Letter of recommendation of the thesis supervisor

Deadline (full application): permanent call
Acceptance: permanent call

The site was built with Mobirise