Keywords: Adaptive Control, Predictive Control, System Identification and Modeling
Abstract: We study output reference tracking for unknown continuous-time systems with arbitrary relative degree. The control objective is to keep the tracking error within predefined time-varying bounds while measurement data is only available at discrete sampling times. To achieve the control objective, we propose a two-component controller. One part is a recently developed sampled-data zero-order hold controller, which achieves reference tracking within prescribed error bounds. To further improve the control signal, we explore the system dynamics via input-output data, and include as the second component a data-driven MPC scheme based on Willems et al.’s fundamental lemma. This combination yields significantly improved input signals as illustrated by a numerical example.

Keywords: System Identification and Modeling, Signal Processing, Robotics
Abstract: This paper presents new viewpoints to solve predicative maintenance problems for arbitrarily rigid-body mechanical systems. Reliable predictions for changes in physical parameters are highly dependent on system model accuracy. Hence, costs for physical system modeling are expensive and time-consuming. To overcome this problem, we replaced the modeling process with a full measurement based procedure in conjunction with radial basis functions (RBF) approximation to create the system model. Subsequently, information theory is employed to generate a meaningful metric quantity that reports about the changes in physical parameters. A rigid-body mechanical system is used as an example to verify the presented theory.

Keywords: System Identification and Modeling, Modeling, Simulation and CAD Tools, Fault Diagnosis and Fault Tolerant Control
Abstract: Strain wave gears (SWG) are commonly used as speed reducers in robot joints. The output torque of these gears can be determined without external measurement equipment by evaluating the signals of strain gauge sensors mounted on the flex spline. This paper presents a simulation chain to compute these sensor signals with the intention to provide a model based approach to study faults in SWGs. It consists of a model describing the dynamic behavior of the gear, a space truss to compute the deformation of the flex spline, and a model for strain gauge sensors mounted on the flex spline. The comparison to experimental data shows the validity of the proposed approach. Moreover, selected faults are injected into the simulation chain and their impact is discussed.

Keywords: Control Systems Design, Robotics, System Identification and Modeling
Abstract: his paper investigates the control of an electric vehicle by an human operator and the techniques needed to compensate for the operator's control deficiencies. Several classes of human operators, of various ages and health conditions, are analyzed. It is shown that the human operator can be characterized by an integrator model with the fractional exponent β ∈[0.8-1.6]. Particular attention is paid to people with Parkinson's disability for whom the exponent β tends to values of 1.5. The fractional model of the human machine system is determined. Control systems based on the use of observers for the estimation of human state components and hierarchical control systems are studied. The stability conditions are obtained by Lyapunov techniques. Experiments on a wheelchair driven by a person with Parkinson's disability confirm the correctness of the proposed solutions.

Keywords: Modeling, Simulation and CAD Tools, Distributed Systems
Abstract: Conventional power converters with lumped elements are widely used today. Distributed converter circuits have been known for a number of years, but are hardly used outside of radio frequency applications. These alternative converters may have qualitatively different properties. We model a distributed boost converter and compare its operation with a conventional boost converter.

Keywords: Software Engineering, Theory of Algorithms, Other Topics
Abstract: The most crucial classes in an object-oriented software system are known as key classes. They serve as the backbone of the software system and are typically documented in project documentation or executive summaries. Key classes play a vital role in software reengineering tasks, particularly when there is a lack of or inadequate documentation for the project. Therefore, the detection of key classes becomes a significant concern.

Various methods have been reported for detecting key classes, including static code analysis, dynamic code analysis, and machine learning. In a static code analysis approach, a weighted graph model is employed to represent the relationships between classes, such as inheritance, interface realization, field usage, and return type usage. By applying a link analysis algorithm on the weighted graph model, the classes are ranked based on their importance, enabling the identification of the top 20 most critical classes.

To evaluate the significance of class relations, we generate all possible combinations of weights for two pairs (1,2) and (1,5) for each class relation during the construction of the weighted graph. Our approach is tested on 14 analyzed systems of varying sizes, encompassing different domains like games, web servers, utilities, and more.

Keywords: Intelligent Systems, Cloud Computing
Abstract: This paper explores the evolution from traditional to modern methodologies for the development of intelligent products. The traditional linear and sequential approach, often characterized by limited toolsets and hardware-centric programming, fails to address the complexities and dynamic requirements of today's digital ecosystems. To bridge this gap, we propose an ecosystem-centric approach to intelligent product development that emphasizes scalability, flexibility, and integration of advanced technologies. Our approach fosters the use of artificial intelligence, machine learning, and advanced data analytics techniques, providing a model for creating scalable and adaptive solutions. We also delve into the necessity of DevSecOps methodology, incorporating cloud-based solutions such as Integration PaaS services, and embracing IoT technologies. This paper contributes by highlighting the benefits of this approach, including improved functionality, reduced development lifecycle, enhanced supply strategy, and cost savings.

Keywords: Optimization, Software Engineering, Intelligent Systems
Abstract: The purpose of this study is to design and build an agile solution, able to adapt to as many practical situations as possible from the spectrum of optimizing distribution routes in the transport of goods. These optimizations are increasingly needed today, in the context of energy and climate crises. Making the use of electricity and fossil fuels more efficient means using a smaller amount of raw material to produce the same result. The study resumes the discussion of the classic problem of the traveler’s commissary as the main starting point in optimizing the distribution routes, over which new functionalities are added to develop an agile algorithm, capable of adapting in as many practical scenarios as possible. Several types of algorithms, such as CVRP, VRPP, VRPPD, VRPTW, etc., are reviewed., explaining a graphical, easy-to-understand manner that clearly shows the difference in approach and the purpose for which they can be used. Reducing distances travelled by optimally grouping points of interest can bring substantial savings. As the study shows, the monthly costs for each vehicle in the fleet are significant, so a reduction in personnel entails a reduction in costs. The work focuses on defining many functionalities to achieve as many existing objectives as possible within any distribution company: minimizing the number of vehicles needed, minimizing transport costs, increasing customer satisfaction, and avoiding penalties. The structure of the paper includes the introduction, the current state of knowledge in the field, the methodology of the research with the presentation of the characteristics of the projected solution, the highlighting of the agile dimension of the proposed solution, the conclusions, and the most important research directions.

Keywords: Biologically Inspired Systems, Biomedical Engineering, Optimization
Abstract: The timing of cellular events is often determined by the instant when the level of a chemical species crosses a threshold from an initial value of zero. To attenuate the inherent random fluctuations in this first-passage time, the production rate can be a function of the current species level by implementing a feedback or self-regulation. In this study, we aim to identify the self-regulation strategy that minimizes timing noise around a specified mean first-passage time. Previous results showed that no-feedback is the optimal strategy in the absence of molecule degradation. Based on this result, we study the case where the species may undergo degradation. First, we set the feedback to be a linear function of the number of molecules. We derive analytical formulas concluding that the optimal strategy involves positive feedback, whereby the production rate increases with the number of molecules. When feedback constraints are removed, our numerical methods suggest that a mixed self-regulation scheme is optimal. This strategy involves an initial positive feedback phase, followed by a sharp transition to a negative feedback scheme in proximity to the threshold. In summary, this research quantifies the fundamental limits of timing noise reduction, considering different strategies of self-regulation in situations where time-keeping species may be subjected to degradation.

Keywords: Internet of Things, Embedded Systems, Communication Systems
Abstract: As renewable energy becomes more prevalent and the Internet of Things (IoT) industry assumes its rightful place in the global technology cycle, research interest in green energy and the possibility of embedded-technology-based wearable devices will increase. The current work focuses on kinetic energy harvesting using footwear mechanisms to generate renewable energy. Human motion is utilized as an innovative method for generating kinetic energy through piezoelectric sensors and recharging batteries for any purpose. The developed system will include a NodeMCU ESP32 (Node Microcontroller Control Unit) development board as a microcontroller to make energy harvesting manageable. The developed prototype will include a solar panel voltage source, making the microcontroller's dynamic charging possible. Furthermore, an IoT interface will monitor the embedded system, allowing the user to check the battery percentage of the piezoelectric-charged battery and the number of successful steps. The connection between the hardware and software components will be established using Bluetooth connectivity. The experimental study entangles multiple essential purposes, including research, health support, environmental protection, and the use of green energy generation techniques.

Keywords: Cyber - Security, Signal Processing, Machine Learning
Abstract: In the current context in which we live, after the health crisis caused by the Sars-CoV-2 virus, preserving data privacy is a sensitive topic and of major importance for everyone. Confidential information has become much more valuable than in the past and has practically become a currency of exchange. This paper proposes a solution for securing X-ray medical records based on an artificial intelligence algorithm that uses convolutional neural networks to transfer the style of a reference image and the MPSA steganographic algorithm to hide the digital X-ray. Furthermore, the paper presents the BeProtected desktop application and the results analysis in terms in terms of imperceptibility and quality of the recovered message.

Keywords: Machine Learning, Computer Vision, Cyber - Security
Abstract: Machine Learning has gained widespread applications, especially in the domain of image recognition due to the high performance of algorithms and models. Nonetheless, the potential threat posed by adversarial samples has impeded their widespread adoption, especially in safety-critical applications. In such cases, the model's final performance is significantly compromised due to the presence of adversarial perturbations that are invisible to human perception. This paper introduces a novel pre-processing solution for protecting image datasets against adversarial attacks such as the Fast Sign Gradient Method or other similar attacks. The proposed method involves utilizing high JPEG compression for removing the adversarial perturbations of the dataset and applying a Targeted One Pixel Attack with the aim of recovering the true class of the images after the compression process. The proposed approach results in a highly secured and clean dataset maintaining a high classification accuracy. The approach was tested on CIFAR10 dataset. The results have shown that the misclassification rate after the FGSM attack was significantly reduced from 89.3% to 12.5% using this method for the compression quality of 20 on a subset of 1000 images.

Keywords: Cyber - Security, Computer Science Education and Training
Abstract: With the increasing importance of cybersecurity in our digital age, cybersecurity competitions have become a popular way to test the incident response abilities of participants. This paper proposes a network architecture to test the vulnerabilities that may appear in a Red and Blue cybersecurity competition. The network architecture is designed to simulate a real-world cyber-attack scenario and includes a router, a core system, and multiple subnets representing different teams. Each subnet contains vulnerable systems that must be protected by the teams, who must also launch attacks on the other teams to identify flags. The architecture presents several challenges, including the large number of rules required for router configuration, which prohibit direct access to opposing teams' virtual machines (VMs) and restrict/block access to specific phases of the competition. The VMs used in the competition include various vulnerabilities related to cryptocurrency wallet operation, medical clinics, chat services, X-ray clinics, SCADA communication protocols, and industrial power plants. Overall, the network architecture and VMs used in this competition provide a challenging and realistic scenario for participants to test their skills in identifying and mitigating cybersecurity threats.

Keywords: Cyber - Security, Communication Systems
Abstract: The analysis of AES (Advanced Encryption Standard) including the byte substitution method, the permutation stage made by shifting the data matrix rows and the column mixing step, shows some of the weak points of this algorithm. We propose enhancing some features of AES by the mean of more powerful arrays and polynomials that make the algorithm more robust against various cryptographic attacks including the differential one. We analyze the algorithm over the Galois Field with 256 integer values, denoted as GF (256), and we propose to use Costas arrays for byte substitution and permutation, and a polynomial with primitive roots for mixing the columns of the data matrix, getting an enhanced version of the encryption algorithm, with better features and an increased degree of non-linearity.

Keywords: Cyber - Security
Abstract: Cybersecurity is vital in our world that uses the public Internet to communicate in every field of activity. In this paper, we present a concise overview of a Security Operation Center (SOC) implementation that combines open source components and operates as a service (SOCaaS). The focus is on the case study of Security Onion, an open source security-monitoring platform, to highlight the effectiveness of this approach. The paper highlights the architecture, key features, and benefits of the SOCaaS model with open source components, emphasizing the integration of Security Onion's intrusion detection, network security monitoring, and log management capabilities. The case study demonstrates the practicality and scalability of Security Onion within the SOCaaS framework. It also addresses challenges and considerations, such as resource requirements, skill gaps, and data privacy. The paper offers insights into how organizations can leverage open source tools like Security Onion to enhance security operations while maintaining control and optimizing costs.