Click here to see the program of SYSTOL'19: https://controls.papercept.net/conferences/conferences/SYSTOL19/program/
Plenary lectures are summarized below.
Recent years have witnessed a growing interest in the field of smart structures. Smart structures are advanced structures with self-capabilities that are emerging as a promoting way to entail innovations beyond the ingenious design of final products. Ongoing programs on their development open new fascinating frontiers for applied research and engineering products. In aeronautic industry, composite materials are increasingly used due to their strength properties. Because of its multilayer structure, they are inherently suitable to host smart materials. Indeed, embedded sensors and actuators could be incorporated permanently into the composite as a smart layer and hence becoming part of the structure.
For aircraft industries, they have come up with the concept of Multifunctional Aircraft Structure (MAS). The principle is to take advantage of new materials to integrate airframe structure with functional systems. The structure has the ability to respond to changes due to environmental conditions and to perform a number of tasks such as Structural Health Management (SHM). SHM is the most form of smartness that is studied. It involves multidisciplinary fields ranging from material, structure, signal processing, data mining, fracture mechanics, fatigue life analysis and more. It aims to detect, localize and evaluate the severity of damages.
As structures age, or undergo fatigue loads, the possibility of failure increases, which may significantly jeopardize operations and safety without timely awareness. Detecting, localizing and monitoring damage in composite structures is a topic of growing interest. The aim of this talk is to review fundamental principles and new developments in SHM. The concept of Damage Tolerant Active Control (DTAC) that allow accommodation of damage by active control methods will also be addressed.
This talk begins with a high-level introduction to novel control-theoretic approaches for the security of control systems. It takes a risk-based approach to the problem and presents a model framework that allows us to introduce and relate many of the recent contributions to the area.
A specific security-related problem is then considered: malicious adversaries that corrupt the control and output signals, with the objective to induce a significant impact on the system performance, while remaining undetected. This attack scenario will be examined more closely from the perspective of a controller / detection filter design problem.
In particular, this talk will draw connections to more conventional approaches, such as optimal robust control and optimal fault detection filter design, and point to key distinguishing features between the fault-tolerance and the secure control perspectives.
Collaborative control of distributed Unmanned Aerial Systems (UASs) assume a tessellation of the free-fly volume into areas of responsibility for each UAS-member. If the UAS has no accurate knowledge of its state (altitude and attitude), the Guaranteed Voronoi (GV) tessellation can be employed. Each UAS is assumed to be informed the uncertain state vector of its Delaunay neighbors and computes in a decentralized manner its GV-cell. Each UAS member is equipped with a Pan-Tilt-Zoom camera and the derived control law computes its altitude and the PTZ-parameters. The overall swarm’s robustness against failures of individual UAS members is examined in this talk. These failures can be due to lack of information exchange between the neighboring UASs and/or actuator failure in each UAS. Simulation studies will be offered to illustrate the effectiveness of this distributed collaborative control law.
Despite the significant advances of SCADA systems for transportation of fluids, important tasks associated with real-time modeling, instrumentation, control and monitoring of pipelines, have been only partially solved. In particular, depending on the flow peculiarities and conduct configurations diverse fault diagnosis methods have been proposed. The main problem is the selection of simple computational models with high sensitivity to the faults, taking into account the reduced number of available hardware sensors and the diverse products that are transported. In this presentation, we share our experience related to the importance of the computational model selection and the excitation signals to design a satisfactory fault diagnosis system. Based on experimental examples of one and two-phase flows, the relevance of physical considerations in the formulation of an FDI task is discussed.