Joseph Yame

On Bumps and Reduction of Switching Transients in Multicontroller Systems

This paper is concerned with the realization and implementation of multicontroller systems, consisting of several linear controllers, subject to the bump phenomenon which occurs when switching between one controller acting in closed loop and another controller in the set of “offline” controllers waiting to take over the control loop. Based on a deep characterization of the bump phenomenon, the paper gives a novel and simple parameterization of such set of linear controllers, possibly having different state dimensions, to cope with bumps and their undesirable transients in switched-mode systems. The proposed technique is based on a non minimal state-space representation allowing a common memory and a unique dynamics shared by all controllers in that set. It also makes each initially open-loop unstable controller run in a stable way regardless of whether that controller is connected to the controlled process.

Performance based switching for fault tolerant control

Abstract This paper presents an approach to active fault-tolerant control (FTC) based on online closed-loop performance monitoring and experimental data without an explicit fault detection and isolation algorithm. The proposed approach sets forth the theory of unfalsified control to achieve a real-time control reconfiguration which is consistent with the performance objectives for a class of system faults. The obtained FTC scheme is an alternative or might be a complement to FTC methods based on explicit fault detectors and provides a precise and direct characterization of suitable real-time controllers for reconfiguration.

A fault accommodation strategy based on closed-loop performance monitoring

This paper presents a new approach to fault-tolerant control (FTC) based on control performance monitoring and real-time plant data without explicit fault detection and isolation algorithms. The approach sets forth the theory of unfalsified control to achieve a real-time reconfiguration of control laws which are consistent with the performance objectives for a class of system faults. The obtained FTC scheme may be viewed as an alternative or a complement to classical FTC structures based on explicit fault detectors and it provides a precise and direct characterization of suitable real-time controllers for fault accommodation.

An Infinite-Dimensional Discrete-Time Representation for Periodic Systems

Abstract This note deals with the representation of continuous-time periodic systems as discrete-time invariant systems in state-space form using the lifting technique. More precisely, we derive an infinite-dimensional state equation on a Hilbert space for continuous-time periodically time-varying systems and characterize the structure of the spectrum of the discrete semigroup generator.

A self-conditioned implementation of switching controllers for smooth transition in multimode systems

This paper deals with an important practical control problem in systems that can switch between different operating modes. We presents a self-conditioned implementation of any two-degree-of-freedom (TDOF) controller which allows a smooth transition when switching between a controller acting in closed-loop and another controller in a bank of idle controllers waiting to take over the control loop. Here, the concept underlying the smoothness of mode transition is that of conditioned transfer which is used to mean that after switching, good tracking performance is ensured albeit a jump at plant input may occur. This notion of conditioned transfer is different from the classical notion of bumpless transfer which refers to the requirement that no bump occurs at the plant input when switching between different controllers. The main feature of our study is that the switching among controllers is driven by the changing dynamics of the controlled-plant and moreover the technique is model-free. This is in contrast with most works reported in the literature on bumpless switching, where switching occurs between controllers driving a plant with fixed dynamics.

On stabilization and spectrum assignment in periodic systems by sampled output feedback controllers

With reference to linear periodically time-varying (LPTV) continuous-time systems, this note points out the fact that the stabilization of such systems by periodic feedback gains with sampled state or output does not imply the relocation of the original characteristic exponents of the LPTV systems as stated previously in the literature.

Representation and stabilizability of continuous-time systems under digital controls

A lifted representation of sampled-data systems is derived on an infinite-dimensional state space; and in this setting their mathematical structure is described and stabilizability questions are reviewed.

Dual adaptive control of stochastic systems via information theory

A new suboptimal dual controller is presented for a class of linear systems with unknown stochastic parameters. The dual property of the controller is obtained through the minimization of a cost function which includes a measure of information about the system.

Conventional structure of a simple class of Takagi-Sugeno PI-type controllers

This paper reveals the conventional structure of a class of Takagi-Sugeno fuzzy PI controllers. This conventional representation provides an insight and understanding on how such fuzzy controllers work and alternatively it provides a mean for the implementation of fuzzy controllers by classical control modules. For the process field engineer, the conventional representation might be of great importance for tuning and maintenance purpose of this class of fuzzy PI controllers.

Controllability in continuous-time systems with digital controllers

The purpose of this paper is the study of controllability properties of a continuous-time system controlled by a digital controller. The study is carried out in the modern lifting framework in which the continuous-time periodicity of such an overall system, called also a sampled-data system, is converted into time-invariance of an infinite-dimensional input/output system in discrete-time. Different concepts of controllability related directly to the continuous-time behavior of sampled-data systems axe introduced, namely exact, approximate and null controllability. It is shown that the two former notions never occur in sampled-data systems while the latter notion is a generic property for these systems.