Head: Claude Delpha

The Modelisation and Estimation Group constitutes one of both groups (with the Inverse Problem Group) of the Signa's and Statistics team at L2S. It includes 16 facuty staff members (8 from CentraleSupelec, 6 from University of Paris-Sud and 2 from the National Center for Scientific Research). A full list of our staff is available here.

Our research interests about information/data sciences include modelisation, representation, estimation, and decision.

We are involved in the Signal Seminar of University Paris-Saclay, the aim of which is to welcome recognized researchers, but also PhD students and post-docs, around the field of signal processing and its applications.

We are mainly interested about methodological approaches with a particular focus on probability and statistics, algebra and optimization. In terms of applications, we propose solutions in various domains such as: defense, biomedical engineering, energy, finance, statistical physics and wireless communications.

More specificaly, our skills include:
- Parametric and non-parametric estimation, robust estimation
- Strongly correlated time series (with potentialy missing data)
- Random matrix theory
- Statistical Inference for Large Scale Data
- Graph theory
- Machine learning
- Stochastic simulation methods
- Multivariate statistics
- Multi-linear algebra and tensor signal processing
- Stochastic differential systems
- Information measures
- Design of experiments
- Time-frequency analysis

Concerning the application fields for which we develop these tools, one can cite:
- Radar signal processing and array processing
- Hyperspectral imaging
- Renewable energy: day-ahead production forecast with time series
- Fault detection and diagnosis for complex systems monitoring
- Health: genetic imaging, biological signals analysis (ECG, heart monitoring)
- Electric load monitoring: identification and filtering of stochastic differential systems
- Wireless communication systems: control and optimization of stochastic differential systems, MIMO networks performance analysis with large random matrices.

Some applications/examples are available here.

We have various academic collaborations in France and around the world. One can cite: Brain and Spine Institute, La pitié Salpétrière Hospital, Paris, Pontificia Universidad Católica de Chile, University of Tulane, USA, Technion Israel, etc.

Several members of our group are involved in academic societies: 3 associates editors for IEEE Transactions on Signal Processing, 1 associate editor for the Chilean Journal of Statistics, 1 member of the technical committee Signal Processing Theory and Method from IEEE, 1 member of the technical committee Sensors Array Multichannel from IEEE, 1 chair and 1 member of the special area team Theoretical and Methodological Trends in Signal Processing from EURASIP, 1 member of the special area team Signal Processing for Multisensor Systems from EURASIP.