In this paper, a fully model-free architecture for vertical stabilization of thermonuclear plasmas in tokamak experimental reactors is presented. For the first time, an Extremum Seeking control algorithm is combined with neural networks to estimate the Lyapunov function to be minimized, resulting in a fully data-driven control architecture. The performance of different neural networks are compared. Specifically, Multilayer Perceptrons and Extreme Learning Machines are considered. The proposed architecture is tested in simulation to show that it can counteract relevant plasma disturbances, resulting in a significant improvement in terms of the achievable operative space compared to the Extremum Seeking algorithm, which still relies on model-based cost estimator. • An innovative fully model-free approach to the Vertical Stabilization problem. • Extremum Seeking combined with a neural network estimation of the cost function. • Control gain event-driven adaptive logic for efficient power supply management. • Comparison of performance and computational complexity of neural networks. • Enlarged operative space without tailoring the gains for specific scenarios.

In this article, the control approach based on structured input–output finite-time stability (IO-FTS) is extended to tackle the reference tracking problem. IO-FTS was originally introduced to deal with the disturbance rejection problem. By applying the finite-time stability control to a properly augmented system, we show that it is possible to enforce a set of specific requirements on the response of the closed-loop system during the transients, taking also into account saturation constraints on the actuators. Adding IO-FTS constraints to classic-state-feedback control law leads to a feasibility problem with bilinear matrix inequality (BMI) constraints. Herein, we show how the original BMI problem can be relaxed to a linear matrix inequality (LMI) one, which comes at a price of more conservatism, but turns out to be computationally more efficient. To prove the effectiveness of the proposed approach, we consider the case of the longitudinal control of a missile.

Diagnosability of faults in discrete event systems modeled with Petri nets can be assessed either via graph-based techniques (also called diagnoser, verifier/twin-plant based techniques), or via the solution of optimization problems. The approaches that belong to the former class are based on the analysis of the net reachability or coverability graphs (or of a more compact version of them). The latter approach exploits the mathematical representation of the net itself to specify and solve optimization problems, which are usually expressed as integer linear programming (ILP) problems. In this paper we exploit the railway Petri net model originally proposed in [16], and extended in [14] to be used as a benchmark for diagnosability analysis, to assess the efficiency of the approach based on the solution of ILP problems proposed in [3]. In order to show the effectiveness of the proposed technique, a comparison with a graph-based approach for analyzing diagnosability is also presented.

This paper aims at defining a model-driven approach for the diagnosability analysis of discrete event systems (DES). The proposed approach can be adopted during the design of modern control systems, in which many sensors and actuators are employed and the diagnosability of faults within a certain delay could be an issue. The proposal represents a first step towards an automatic model-driven process which derive formal models from a complete high-level specification of DESs. The specification activity of our approach relies on the Dynamic STate Machine (DSTM) formalism, a new language that extends state machines with dynamic instantiation, interrupts and asynchronous communication. The paper will describe how we can automatically derive Petri net and Promela models from the high-level DSTM specification. The former model can be used to apply diagnosability analysis approaches proposed in the DES community, while the latter can be used to apply model checking techniques. An application of the proposed model-driven approach is described by deriving both a PN and a Promela model for the well-known railway level crossing benchmark.

This paper presents the notion of real-time containers, or rt-cases, conceived as the convergence of software container technologies, such as Linux Containers and/or Docker, and real-time operating systems. The idea is to allow critical containers, characterized by stringent timeliness and reliability requirements, to cohabit with traditional non real-time containers on the same hardware. The approach allows to keep the advantages of real-time virtualization, largely adopted in the industry, while reducing its inherent scalability limitation when to be applied to large-scale mixed-criticality systems. The paper provides a reference architecture scheme for implementing the real-time container concept on top on a patched real-time Linux kernel, and it overviews the challenges to be faced to implement the rt-case vision.

Model-driven design is recently gaining a wide spreading in different fields, such as design of mechatronics and embedded systems. Approaches based on either UML or SysML permit to efficiently manage the design of complex systems in different areas. In this paper a SysML modeling approach is proposed for the design of real-time applications based on the MARTe framework. The MARTe framework has been recently adopted for the development of real-time systems in several European experimental fusion reactors. The proposed approach allows to achieve a better standardization of the development cycle and a better documentation of the developed systems. Furthermore, by using model-2-text tools it is possible to automatically generate part of the real-time executable code and the application configuration file. The proposed approach has been applied for the modeling of the control system for the Frascati Tokamak Upgrade.

This paper deals with the problem of identifying a Petri net system given an observed sequence of events generated by it and an observed sequence of output vectors associated to the marking of the measurable places. The problem is not new in the literature. The original contribution of this work consists in the use of the timing information associated to the net so as to improve its identification. The fact that a transition has not fired for a time larger than its expected delay is here exploited to obtain a list of counterexamples, i.e. the strings that does not belong to the language, even if a whole language is not known. This accelerates the identification procedure. The problem of identifying a ¿-free labeled Petri net system is here considered, that is the identification of a net where a label may be associated to more than one transition, and no transition may by labelled with the empty string.