Grigore Stamatescu

Grigore Stamatescu

Professor at UPB

Teaching

Description: The course aims to present and transfer knowledge on the formalization, design, implementation and testing of modern artificial intelligence systems. Theoretical aspects are presented that correspond to the representation of knowledge and the treatment of uncertainty through probabilistic approaches. Bayesian models and networks as well as simple and complex decision algorithms based on Markov processes are discussed. Automatic planning methods based on heuristics and hierarchical planning leading to optimization and efficiency of resource use are classified and compared. Advanced neural network architectures for information encoding are explored: graph-based neural networks, autoencoders, and generative models. Applications of artificial intelligence in high value-added fields such as natural language processing, robotics and artificial vision are presented through representative examples. Aspects relevant to the standardization and regulation of artificial intelligence systems in critical applications are also discussed. Aspects of philosophy, ethics and safety in artificial intelligence also play an important role in the widespread adoption of these solutions, with a significant social and economic impact.

Keywords: artificial intelligence, decision algorithms, automated planning, ai applications, ai ethics, explainability and safety

Description: The main objective of the Intelligent Measurement Systems course is to present the main design concepts and applications of networked embedded systems in environmental, urban, industry and energy monitoring, as fundamental building blocks of the new Internet of Things (IoT) and Cyberphysical Systems (CPS) concepts. Node-level aspects: sensor selection, energy issues and local data processing, and network-level aspects: wireless communication, localization, mobility, routing and distributed algorithms, are addressed in detail. The social implications of the widespread adoption of these systems are also discussed. Practical simulation applications are developed using the CubCarbon hybrid IoT environment as well as experimental implementations on a dedicated WSN test bed.

Keywords: networked embedded systems, sensors, distributed data processing, wireless sensor networks, internet of things

Description: The course presents the issues of ensuring the cybersecurity of industrial control systems (ICS), exposed to a growing number of cyber threats and vulnerabilities, which can have major consequences in essential sectors for society. The main methodologies, specific terminology and standardized approaches to risk management and the implementation of appropriate technical controls for industrial automation equipment and networks are presented. A comparative analysis is also carried out for the comparative characterization of cyber security for IT/OT systems. Typical Vulnerability Analysis Patterns (CVSS) and Cyber Security Incident Reporting and Information Sharing (MISP) methods are defined. Robust configuration and parameterization methods are discussed using specialized software tools for securing industrial equipment. The current international standardization (ISO27001, IEC62443) and regulatory (Cybersecurity Act, NIS2, Cyber Resilience Act) framework is also discussed, along with the associated technical resources to ensure compliance.

Keywords: ics cybersecurity, ot security, cyber-threat intelligence, misp, standardisation, regulation

Description: The course on Control Engineering (Regelungstechnik) consists of an introduction to the theory of linear dynamic systems and the design of control systems, with modeling and simulation applications and the presentation of pilot installations. Topics related to the mathematical modeling of physical systems and the study of their behavior in the time and frequency domains are addressed, based on test signals. The stability of continuous linear systems is defined and evaluated. Methods of analysis and synthesis of the regulation and design loops of P- / PI- / PID type regulators are presented, including the optimization of tuning parameters, with applicability in the management of industrial processes and intelligent embedded systems. Modern software tools are used e.g. development environment for MATLAB / Simulink technical computing and interactive teaching methods.

Keywords: modelling and simulation, time analysis, frequency analysis, stability, automatic control systems, matlab / simulink

Description: The course on Control Engineering II is a continuation of the discipline of Control Engineering (Regelungstechnik I). The method of analysis and design of systems based on root locus is presented. An introduction is made in the field of state space and an analysis and synthesis of nonlinear systems in the domain of frequency and in state space. The controllability and observability properties of the systems are described and evaluated. Intelligent techniques for control of industrial processes using fuzzy logic, neural networks and genetic algorithms are also presented. The project assignment developed during the course integrates the knowledge gained in a unitary approach, with practical applicability through appropriate software tools: MATLAB Control System Toolbox, SISO Tool, Fuzzy Logic Toolbox.

Keywords: root locus, state space, multivariable system, nonlinear systems, intelligent control systems

Keywords: informatics, programming, python, data structures and algorithms, object oriented programming