Optimal Filtering for Information Estimation

The ongoing integration of information technologies into modern systems is based on an ever-increasing appropriation of advanced Signal Processing methods. The aim of this advanced course is to complete and extend undergraduate lecture courses by exploring the problem of estimating information through filtering, and by presenting optimal resolution methods. It enables engineering students to acquire new tools for signal modeling, analysis and filtering. Particular emphasis will be placed on the practical value of the proposed methods, for example through the use of efficient algorithms or the consideration of filter implementation constraints.

This course is designed for students wishing to continue their studies in Signal Processing, as well as for those only interested in this general topic, which is used in a wide range of applications (telecommunications, energy production and storage, transport of goods and people, aerospace/aeronautics, industry 4.0, medicine, etc.).

Lectures

• Introduction and formalization of the estimation problem by optimal filtering

• Resolution by input/output approach: Wiener filtering

• Resolution by state-space approach: Kalman filtering

• Practical application: tuning, algorithm, implementation

Practical work

• Application through a practical example of a sound signal recording in a noisy environment

Project part

• Realization and evaluation of a complete information estimation system through an application project (e.g. estimating the position of a drone in a constrained environment).