Volume 59, Issue 12 -

Models are commonly derived and assessed with respect to their prediction error/approximation capability — independent from the intended model deployment. This origins from the typically used modeling methods. The second reason is that modeling and control design are often carried out by different specialists. As an exact model can hardly be obtained, available degrees of freedom in modeling should be used to obtain a model that is optimal for the intended application. For model-based control design, the suitability for control design is the primary criterion but not the prediction performance of the model. This insight triggered the development of new approaches in the area of linear system identification. This contribution presents a method for deriving control-oriented Takagi-Sugeno models from physical models. The method is demonstrated on example of the benchmark system “inverted pendulum”.

This paper presents a generalized description of linear transport processes that allows for a derivation of common dynamic properties and a control layout for a wide set of applications. Based on the introduction of the system description of platoons and a proposition of suitable performance criteria a control layout is formulated which reduces the synchronization of a common motion of sub-systems to a stabilization problem of the platoon and optimizes the corresponding performance measures. Furthermore, the influence of communication topology and switching information access among platoon members is investigated. The proposed approach is applied to a vehicle platoon with one leading and three following vehicles and is analyzed through both simulations and physical experiments with model scale vehicles. The results show that the approach is also applicable to physical plants that can only be approximated by a linear platoon model. Zusammenfassung In diesem Beitrag wird eine verallgemeinerte Beschreibung von Transportprozessen vorgestellt, die eine Ableitung gemeinsamer Systemeigenschaften sowie die Auslegung von Regelungen für eine Vielzahl von Prozessen ermöglicht. Auf Basis der eingeführten Systembeschreibung von Kolonnen und der Formulierung geeigneter Gütekriterien wird ein Reglerentwurf vorgeschlagen, der die Synchronisierung einer gemeinsamen Bewegung aller Teilsysteme auf die Stabilixsierung der Kolonne zurückführt und eine optimale Lösung in Bezug auf die formulierten Güteanforderungen darstellt. Des Weiteren wird der Einfluss der Kommunikationstopologie und der schaltenden Änderung von Informationsverfügbarkeit innerhalb der Kolonne untersucht. Der vorgeschlagene Entwurf wird auf das Beispiel einer Fahrzeugkolonne mit einem Führungsfahrzeug und drei Folgefahrzeugen angewendet und durch Simulationen sowie durch Experimente an Modellfahrzeugen analysiert. Die Ergebnisse zeigen, dass sich das Verfahren auch auf physikalische Strecken, deren Dynamik lediglich durch lineare Modelle approximiert werden kann, anwenden lässt.

Stainless steel strips are usually produced on 20-high rolling mills. The involved processes are complicated and multivariable systems with dominant time delays. The most important quality parameters, strip thickness and strip flatness, have to be optimally controlled. The control system should consider the mutual interaction between both controlled variables, in particular for strips of high strength and low thickness. This contribution presents an integrated thickness and flatness control system, designed to consider and compensate the interaction between both quality parameters. A multivariable control approach with decoupling is used, in which the parameters are online adapted by the instantaneous linearization of a physically-based roll-stand model. The implementation of the new controller in an industrial Sendzimir mill shows that it allows the use of actuators at maximum speeds and thus guarantees best control performance for both strip thickness and flatness simultaneously.

An innovative hybrid system structure for continuous control systems with discrete decision-making processes is presented. The functionality is demonstrated on a highly automated driving system on freeways and on the emergency stop assistant. Due to the fact that the robustness will be a determining factor for future usage of these systems, the presented structure focuses on this feature.