Section: Application Domains
Industry and society: i-PID for industry and society, mechatronics (Safran)
Industry is keen on simple and powerful controllers. The tuning simplicity of the classical PID controller explains its omnipresence in industrial control systems, although its performances drop when the working conditions change. AL.I.E.N SAS was created in 2011 as a spin-off of the Inria project ALIEN, which gave rise to Non-A, working on algebraic estimation and i-PID controller (i.e., using algebraic estimation of the perturbations and apply a simple PID control on some “ultra-local” model). These control technique uses the information contained in the output signal and its estimated derivatives, which can be regarded as “signal-based” controllers. Model-free control technique has been applied in many different domains (electronics, hydroelectric power, etc.).
Recent research is focused on traffic control and biology. The quality of traffic control laws depends on a good knowledge of the highway characteristics, especially the critical density and the free-flow speed, which are unfortunately most difficult to estimate in real time. Therefore, we aim at developing an algorithm which shows the possibility to control the traffic without the knowledge of density and free-flow speed.
A collaboration with the Safran Electronics & Defense company has been developed (CIFRE PhD thesis) on the parametric stabilization of gyrostabilized platforms. To do that, we first aim at developing new symbolic-numeric methods for the standard -loop shaping design problem for models of gyrostabilized platforms in terms of the physical parameters (masses, inertia, etc.) considered as unknown/slowly varying parameters. Using Non-A techniques for the estimation of the physical parameters, we then want to develop new embeddable and adaptive controllers for the robust stabilization of gyrostabilized platforms.