ALIEN Research team
Algebra for Digital Identification and Estimation
- Leader : Michel Fliess
- Research center(s) : CRI Saclay - Île-de-France
- Field : Applied Mathematics, Computation and Simulation
- Theme : Modeling, Optimization, and Control of Dynamic Systems
- Partner(s) : Ecole Polytechnique,CNRS,Ecole Centrale de Lille
- Collaborator(s) : Laboratoire d'informatique de l'école polytechnique (LIX) (UMR7161),Laboratoire d'Automatique, de Génie Informatique et Signal (LAGIS) (UMR8219)
Team presentationIn the fields of control and signal processing, identifying the parameters of a model or estimating non-measured variables are common problems, normally processed within the context of optimisation. The ALIEN project-team, which has been created in June 2004, is developing an algebraic approach that provides explicit formulae. This non-asymptotic character is a significant advantage for real-time applications.
For engineers, a wide variety of information is not directly accessible to measurement. Some parameters (constants of a magnetic machine, delay time in communication, etc.) or internal variables (mechanical torques in a robot, etc.) are unknown. Similarly, more often than not, signals from sensors are distorted and tainted by measurement noises. To control such processes, and to extract information conveyed by the signals, one often has to identify a system and estimate parameters.
The ALIEN project-team is developing an estimation theory, built around differential algebra and operational calculation. It has resulted in relatively simple, rapid algorithms: Solutions are provided by explicit formulae, with straightforward implementation, using standard tools from computational mathematics. Unlike traditional methods, the majority of which pertain to asymptotic statistics, the ALIEN estimators are “non-asymptotic”. In many application sectors, the response time parameter is crucial. Using this approach, computations are performed as the application is running: The “real-time” computing is targeted, as opposed to processing that occurs after the event.
Application fields for this development are plentiful: Today the team is working on the control of very high precision systems as well as NCS (networked control systems). Applications in nanobiology are also under study.
Research themesThe identification of linear systems, in the sense of automatic control, is benefiting from the algebraic module theory and from operational calculus. It permits perhaps for the first time to work in real time, i.e., to simultaneously identify and control, a fact which is often indispensable in practice. For the nonlinear generalisation we are solving a long-standing problem, i.e., the estimation of the derivatives of various order of a noisy signal, in a way which is easy to implement. Works in progress demonstrate that we are not only able to identify the poorly known parameters, but also to estimate the state: our renewed perspectives yield for the first time a systematic procedure for obtaining non-linear observers. Diagnosis, i.e., fault detection and isolation, and their accommodation, may be accomplished thanks to those techniques in closed loop, i.e., by controlling and identifying the system. Various concrete illustrations have been already tested successfully (control of an electric plant at the CINVESTAV, Mexico City, Mexico; control of a rotating machine at the Technical University of Dresden, Germany). The on-going control-oriented applications concern: (1) The control of active magnet bearings and other levitation-based systems (collaboration with the Technical University of Dresden, Germany); (2) The design of actuators based on shape-memory materials (collaboration with the University of Brest); (3) the control high quality machining robots, requiring a rapid response upon which the precision and surface quality on the parts obtained depend (collaboration with ENSAM Lille and Dynalog Europe).
In what concerns the signal processing, similar methods yield answers to denoising, to the detection of abrupt changes, to demodulation, blind equalization and compression, even for transient signals in a quite noisy environment. Two patents related to those techniques, which are of utmost industrial significance, are pending. The extension to image and video signals yields remarkable results for denoising, compression, edge and motion detection.
Last, both control and signal aspects lead to very interesting applictions in the domain of nanovirology. ALIEN leads an active collaboration with the Laboratory of Virology (CHRU, University of Lille 2) and the LNE (Laboratoire National de métrologie et d'Essais, Trappes). It is aimed at increasing the precision and rapidity of atomic force microscopes (FSA) so to improve significantly the knowledge in fundamental virology and the potential of AFMs in biology. To our best knowledge, the imaging we obtained in 2007 for a Coxsackievirus B4 under physiological temperature conditions constitutes a European première.