L'objectif de cette thèse consiste à étudier et à développer des modèles prédictifs de production d'énergie solaire en fonction des paramètres météo. Ces modèles devront être implémentés dans le contexte de la plateforme de gestion de centrales photovoltaïques et serviront aux application suivantes :

• pour le court terme : la mise en place d'une détection d'anomalie de la production électrique (comparaison production simulée - production réelle)
• pour le long terme : un service de prévision à quelques jours de la production électrique globale sur un ensemble de centrales

Contacts

http://solarnet.fr/

 

 

PhD Position

• Doctoral school: EDITE (Telecom ParisTech)
• Co-direction ISEP (SITe) - Telecom ParisTech (TSI)
  

Research Context

The MCUBE project consists of providing a generic M2M (Machine to Machine) system for multimedia applications, i.e. involving sound and image analysis. The goal is to use standard video and sound capture devices for supervision applications in the Agricultural (Cap2020 partner) and Photovoltaic (SolarNet partner) domains. For example, image analysis can be used to follow crops growth in the fields, or to check the state of solar panels for maintenance. Sound analysis may be used to count insects in the field, to detect intrusions in the plants, etc. The MCUBE project will provide a service for finding and deploying such algorithms on M2M gateways (WebDyn partner). It will have to take into account typical M2M constraints, such as energy consumption, communication bandwidth and cost, relatively small memory and disk resources, etc. These environmental constraints may vary depending on the deployment target and the system will need to take this variability into account too.

In order to set up an operational platform, the PhD student will work with the ISEP researchers in Software Engineering and Signal Processing. The thesis will be co-directed by a researcher from the TSI department of Telecom ParisTech. A research engineer will also be allocated to the project for helping with the implementation and the deployment of the solution.

The PhD student will also have the opportunity to be a teaching assistant in his field of competences.

PhD Topic: Signal and Image processing. Analysis of multimedia data in M2M applications, real-time classification.

The candidates must have a strong background in signal processing, image processing and analysis and hold a MSc or equivalent diploma. Especially, the following skills are required:

• Fourier analysis,
• digital filters,
• wavelet analysis,
• image enhancement,
• image segmentation,
• multiple image registration,
• feature recognition and classification (texture, shape),
• morphological operations,
• image and data fusion,
• object tracking in video sequences.

Some experience with research is a plus, especially for engineers.

English writing and speaking fluency is required. French speaking also.

Application

Please send your resume and cover letters in English to Florence Rossant. A recommendation letter from a professor is a plus.

Location

ISEP, 21 rue d'Assas, 75006 Paris

Annual Gross Wage (estimate)

+/- 25000 euros

Interference alignment: frequency and time-based comparison

Interference alignment for wireless networks is a recent technique for increasing the multiplexing gain on the interference channel. It consists in forcing interfering signals at each receiver into a reduced-dimensional subspace of the received space, so that the receivers can observe an interference-free desired signal. The considered space may be an actual space (time, frequency, physical path) or structural space of the signal.

This training work will study interference alignment in time and frequency. Multiple user access will be either obtained via Time Division Multiple Access (TDMA) or Frequency Division Multiple Access (FDMA). The objective is to characterize, depending on the QoS constraints, users load and traffic load, which multi-access technique is the most efficient to fulfil users QoS constraints.

The simulations will be dynamic, accounting for the users traffic load. New access techniques for interference alignment will be determined.

Keywords:

Interference alignment, radio access, TDMA, FDMA

Requirements:

Telecommunications, Matlab and/or C languages.
Good mathematical level.

References:

• C. Suh and D. Tse, "Interference alignment for cellular networks", in Allerton Conf. Commun., Control and Computing, Sept. 2008.
• K. Gomadan, V.R. Cadambe, S.A Jafar, "Approaching the capacity of wireless network through distributed interference alignment", ArXiv, March 2008.

Contact

Spectrum sensing techniques for cognitive radio

The secondary network must detect the unused spectrum. Detecting primary user sis the most efficient way to detect spectrum holes. Several techniques exist for spectrum sensing. They can be classified into three categories:

• Transmitter detection: cognitive radios must have the capability to determine if a signal from a primary transmitter is locally present in a certain spectrum, there are several approaches proposed:

• Marched filter detection
• Energy detection
• Cyclostationary feature detection

• Interference based detection.

The training work will consist in studying these techniques, and comparing them. More precisely, the advantages and drawbacks (in terms of requested information, complexity, etc.) of each method will be determined. The performances of spectrum sensing techniques will be evaluated through MatLab or C simulations. This characterization should lead to a set of recommendations. The context of application will be a primary WiMAX network, and a secondary network that may be either based on OFDMA, or on CDMA.

Keywords:

Cognitive radio, spectrum sensing, detection

Requirements:

Telecommunications, signal processing, Matlab and/or C languages.

Good mathematical level.

References:

• D. Cabric, S.M. Mishra and R.W. Brodersen, "Implementation issues of spectrum sensing in cognitive radio", in ASILOMAR Conf., 2004.
• A. Sahai and D. Cabric, "Spectrum sensing: fondamental limits and practical challenges", Tutorial, Dyspan Conf., 2005.

Contact