| |
Last updated on October 17, 2024. This conference program is tentative and subject to change
Technical Program for Thursday October 10, 2024
|
ThA1 Regular session, George Enescu |
Add to My Program |
Control Systems and Applications (1) |
|
|
Chair: Popescu, Dan | University of Craiova |
Co-Chair: Burlacu, Adrian | Gheorghe Asachi Technical University of Iasi |
|
09:30-09:50, Paper ThA1.1 | Add to My Program |
Towards Learning-Based Trajectory Tracking Control for a Planetary Exploration Rover: Adaptive Model Predictive Control |
|
Baldauf, Niklas | E: Fs TechHub GmbH |
Lakatos, Kristin | German Aerospace Center (DLR) |
Meinert, Alexander | E: Fs TechHub GmbH |
Turnwald, Alen | E: Fs TechHub GmbH |
Keywords: Adaptive and Robust Control, Predictive Control, Control Applications
Abstract: Automous Robotic vehicles are important for exploring distant planets due to the risks associated with manned missions. However, they face challenges in navigating unknown terrain, requiring a deep understanding of wheel-ground interaction for effective slip estimation and compensation. Conventional methods have proven insufficient in addressing these challenges. The DeLeMIS project showcases AI methods to enhance autonomy in uncertain terrain, focusing on the trajectory tracking performance. This paper presents the development and evaluation of an adaptive model predictive control algorithm within the project, extended by machine-learning techniques. The learning-based algorithm and conventional baseline controllers are implemented and systematically evaluated through a comprehensive test campaign, which is described in our previous publication. This paper gives an overview of the control architecture and the training process, providing detailed insights into the neural network architecture and the results on a prototypic rover hardware.
|
|
09:50-10:10, Paper ThA1.2 | Add to My Program |
Low Level Adaptive Control of Vehicle Longitudinal Dynamics: A Driver Inspired Architecture |
|
Penet, Maxime | Valeo |
Le-Gall, Gaetan | Valeo |
Keywords: Automotive Control Systems, Autonomous Systems, Control Applications
Abstract: This paper deals with the design of a low level controller for acceleration tracking using the vehicle's pedals. Its architecture aims at mimicking the various elements which are involved when a human is driving. The focus is on providing an efficient and easy to deploy solution. This is achieved by using data collected from non tailored driving sessions and requiring basic vehicle's knowledge. This topic takes all its relevance in the context of autonomous driving where the focus is usually on features which provide high level requests such as acceleration. Performances are claimed under the tacit requirements that requests will be accurately tracked by a (real) vehicle. This is however a strong assumption with respects to chassis control challenges, wide deployment concerns and calibration by non experts. From a calibration perspective, the focus is on how to process data collected in a relatively uncontrolled way to automatically generate appropriate parameters. From a feedback design perspective, the challenge is to retain a performing yet easy to interpret controller structure. To illustrate the controller's performances, tests results obtained from real life experiments are presented.
|
|
10:10-10:30, Paper ThA1.3 | Add to My Program |
Side-Slip Compensation in Model Predictive Path Following Control for General-N-Trailer Systems |
|
Koegler, Philipp | Friedrich-Alexander-Universität Erlangen-Nürnberg |
Dahlmann, Julian | Friedrich-Alexander-Univerity Erlangen-Nürnberg |
Graichen, Knut | University Erlangen-Nürnberg (FAU) |
Keywords: Automotive Control Systems, Predictive Control, Autonomous Systems
Abstract: This paper presents a concept for the incorporation of side-slip angles in a model predictive path following controller for multi-trailer vehicles in off-road environments. A common approach for trajectory tracking and path following of so-called general-n-trailer systems is the application of a model predictive controller with an internal kinematic model. In this context it is often assumed that the vehicle operates on a plane surface without slippage. However, when employing vehicles in harsh environments, these assumptions are violated, resulting in a model discrepancy and thus in an impaired tracking accuracy. Through extending the kinematic general-n-trailer system by additional disturbance inputs, a model of low complexity is provided that can still represent the side-slip behavior. Additionally, a method for predicting future side-slip based on the extrapolation of current estimates is proposed. As heavy-duty vehicles are often utilized in repetitive processes, it is further examined how knowledge about the sliding behavior can be acquired over successive runs.
|
|
10:30-10:50, Paper ThA1.4 | Add to My Program |
Human Machine System Control for Persons with Operating Deficiencies |
|
Ivanescu, Mircea | University of Craiova |
Nitulescu, Mircea | University of Craiova |
Keywords: Control Systems Design, Biologically Inspired Systems, Human - Computer Interface
Abstract: The paper focuses on the design of a control system for an operator with Parkinson's diseases driving an electric vehicle. The fractional order model of a person with Parkinson’s disease is discussed. The fractional model is extended to the dynamics of an electric car operated by a person with disabilities. The associated mathematical models are analyzed, insisting on the influence of dead times on driving performance. The effects generated by dead times on the visual perception of motion parameters as well as dead times on the generation of a control decision are discussed, times determined by the incapacity of the operator affected by Parkinson's disease. Methods for approximating the dead time dynamics are proposed and the stability of the human-machine system is investigated, as a hierarchical system, using vector Lyapunov techniques. The theoretical results are verified by numerical simulation.
|
|
10:50-11:10, Paper ThA1.5 | Add to My Program |
Model Predictive Traction Control System Based on the Koopman Operator |
|
Kir Hromatko, Josip | University of Zagreb Faculty of Electrical Engineering and Compu |
Iles, Sandor | University of Zagreb, Faculty of Electrical Engineering and Comp |
Keywords: Predictive Control, Automotive Control Systems, Data - Driven Control
Abstract: Due to their importance, traction control and anti-lock braking systems have become standard equipment in modern vehicles. However, accurate models of tire dynamics are often difficult to obtain and usually include nonlinearities, making their use in control systems challenging. This paper describes a traction control system based on model predictive control and Koopman operator theory, which aims to approximate nonlinear systems with linear ones through a state space transformation. A linear model predictive controller based on the Koopman predictor is compared to a standard nonlinear model predictive controller. Experiments in a high-fidelity vehicle dynamics simulation environment show a comparable reference tracking performance of the two controllers, with a reduced execution time for the proposed Koopman operator-based algorithm, both on a standard PC and embedded hardware.
|
|
11:10-11:30, Paper ThA1.6 | Add to My Program |
Cloud-Based DMPC Simulation for Autonomous Mobile Robot Platooning |
|
Maxim, Anca | "Gheorghe Asachi" Technical University of Iasi |
Pauca, Ovidiu | “Gheorghe Asachi” Technical University of Iasi |
Lazar, Razvan-Gabriel | Gheorghe Asachi Technical University of Iasi |
Amarandei, Cristian-Mihai | "Gheorghe Asachi" Technical University of Iasi, Faculty of Autom |
Caruntu, Constantin-Florin | Gheorghe Asachi Technical University of Iasi |
Keywords: Cloud Computing, Autonomous Systems, Predictive Control
Abstract: Cloud computing can be an useful tool when dealing with spatially distributed resources, which need to be simultaneously managed. In this work, a simulation environment is developed using OpenStack to apply a Distributed Model Predictive Control strategy for a platooning application consisting of multiple autonomous mobile robots. The idea is to develop the entire algorithm by taking advantage of the possibility of creating virtual machines. Thus, each autonomous mobile robot is treated as an individual agent and is modeled and controlled on a virtual machine, whereas the communication between the agents is performed in the cloud, using the TCP/IP protocol standard. The simulation results illustrate that the platooning application deployed in OpenStack outperforms the simulation performed on a single computer.
|
|
ThA2 Regular session, Nicolae Iorga |
Add to My Program |
Biomedical Engineering & Biologically Inspired Systems |
|
|
Chair: Barbu, Marian | Dunarea De Jos University of Galati |
Co-Chair: Angelescu, Nicoleta | Valahia University of Targoviste |
|
09:30-09:50, Paper ThA2.1 | Add to My Program |
Classification of Skin Lesions Using Morpho-Granulometry Features Derived from Color Clusters |
|
Tăbăcaru, Gigi | “Dunarea De Jos” University of Galati |
Moldovanu, Simona | Dunarea De Jos University |
Barbu, Marian | Dunarea De Jos University of Galati |
Keywords: Biomedical Engineering, Machine Learning
Abstract: The skin lesion can be thought of as a biological system, so the morpho-granulometry of significant color clusters found in skin lesions is one of the elements that reproduce in a natural way the structure of the lesion, this novelty is highlighted in this study. Important features of skin lesions can be modulated by fusing neural networks (NN) and machine learning (ML). By choosing the nevus and melanoma classes, the primary goal was accomplished, and three databases were used to test the methodology. The characteristics based on morpho-granulometry allowed for the identification of microstructure within the images, which can be very helpful in characterizing the biological system. Based on random forest (RF) and extreme gradient boosting (XGboost) classifiers, this work aimed to improve the classification performance of important feature selection. The selected features from three free image databases with three NNs were classified. In a binary classification of nevus vs. melanoma, the results showed that the pattern recognition neural network (PRNN), according to the PH2 database, provided an accuracy of 0.923 and an F1-score of 0.876. The classification is interpretable if it is not validated. In our study, the best results were verified with a logistic regression (LR) classifier.
|
|
09:50-10:10, Paper ThA2.2 | Add to My Program |
Koopman Linearization and Optimal Control of Glucose Level |
|
Pintea, Paul-Andrei | Technical University of Cluj-Napoca |
Mihaly, Vlad Mihai | Technical University of Cluj-Napoca |
Susca, Mircea | Technical University of Cluj-Napoca |
Dobra, Petru | Technical University of Cluj |
Keywords: Biomedical Engineering, Nonlinear Systems, Optimization and Optimal Control
Abstract: The Artificial Pancreas Problem (APP) offers a potential framework for Control Engineering studies, specifically in the field of continuous monitoring and actuation to control glucose levels. The models that give a satisfactory level of accuracy are nonlinear by nature however, the standard approach in linear control is to find a linear representation of the model. The current paper proposes a comparison between standard linearization and linearization via the Koopman Operator for an input-affine nonlinear model from insulin intake to glucose level. Each model also has an additive disturbance component. To account for it, the current paper proposes a method of modeling the disturbance based on Gauss Processes. For a meaningful comparison between the considered linear matrix inequality-based controllers (LMI) and linear-quadratic regulators (LQR), the paper introduces the term Glucose Absolute Error (GAE) as an error index adapted for the Insulin-Glucose system.
|
|
10:10-10:30, Paper ThA2.3 | Add to My Program |
Heart Rate Variability-Based Software Analysis of Cardiac Data |
|
Tsanev, Yoan-Aleksandar | Technical University of Varna |
Georgieva-Tsaneva, Galya | Institute of Robotics |
Cheshmedzhiev, Krasimir | Institute of Robotics, Bulgarian Academy of Sciences |
Keywords: Signal Processing, Software Methods and Tools, Information Systems Applications
Abstract: Cardiovascular diseases are the cause of increased mortality in all countries of the world, which explains the deepening of their study and the search for additional means of analyzing cardiac data. The article presents mathematically based methods for the processing and analysis of cardio data that are recorded during the normal life of people. The results of linear (in the time and frequency domain) and non-linear procedures (Poincarè method) as well as graphical methods (Power Spectral Density) for determining heart rate variability in healthy and heart diseased individuals are shown. A statistical analysis was applied to determine the significance of the obtained results when comparing the indicators of the two studied data sets.
|
|
10:30-10:50, Paper ThA2.4 | Add to My Program |
Adaptive Control Design for Blood Glucose Regulation of T1DM Patient: An LMI Framework |
|
Anamika, Anamika | NIT Silchar |
Datta, Bidipta | NIT Silchar |
Dey, Rajeeb | National Institute of Technology Silchar |
Nath, Anirudh | GE Research |
Gavrila, Tania Bogdana | UAV Arad |
Balas, Valentina Emilia | Aurel Vlaicu University of Arad |
Keywords: Biomedical Engineering, Adaptive and Robust Control, Nonlinear Systems
Abstract: —This paper deals with the design of an MRACbased adaptive control law for blood glucose regulation of T1DM patients. Parametric uncertainty in T1DM patients is one of the primary issues in designing adaptive control laws. By designing an adaptive controller in the MRAC framework, the risk of hypoglycemia is eliminated. Extensive simulation experiments are used to examine the closed-loop response of plasma glucose concentration and external insulin infusion rate for a wide range of model parameter variations.
|
|
10:50-11:10, Paper ThA2.5 | Add to My Program |
Neural Network Based Predictive Model for Bevel-Tip Needle Deflection in Percutaneous Interventions |
|
Maria Joseph, Felix Orlando | Indian Institute of Technology Roorkee |
Keywords: Instrumentation, Mechatronics, Biomedical Engineering
Abstract: This study employs a neural network-based modeling approach to forecast the deflection of a Brachytherapy needle upon insertion into a phantom tissue. The needle, featuring an asymmetric tip, deviates from its original position within the tissue, impacting the efficiency of the treatment. Addressing this issue requires a focused effort to minimize deflection. In this research, needle deflection is predicted using a multilayer perceptron technique, a data-driven analysis applicable to various needle-tissue combinations. The model conceptualizes the system with a distributed load acting on the slender part surrounded by the tissue and tip force. A notable contribution of this paper is the incorporation of distribution load from raw sensor data to enhance the precision of deflection prediction. The model's efficacy is validated through 5-fold cross-validation. Performance evaluation reveals that the proposed model predicts deflection more accurately than results obtained using the Euler-Bernoulli beam theory.
|
|
11:10-11:30, Paper ThA2.6 | Add to My Program |
Modeling a Viral Infection Process Using Dynamic P Systems and High Level Petri Nets with Object Orientation |
|
Brezovan, Marius | Universitatea Din Craiova |
Stanescu, Liana | University of Craiova |
Keywords: Biologically Inspired Systems, Petri Nets, System Identification and Modeling
Abstract: Modeling has an important role in systems biology because it can provide a better understanding of biological systems. Advances in biological systems have introduced a series of challenges, such as the repetition of components (cells), their variation and hierarchical organization. There is a wide variety of modeling approaches, including Petri nets, Brane calculi, and P-systems. In this paper we propose an approach based on object-oriented Petri nets, which encodes a category of P systems, called dynamic P systems. In addition, we show the validity of the modeling method by modeling a viral infection process with the influenza virus.
|
|
ThA3 Regular session, Mircea Eliade |
Add to My Program |
Neural Networks |
|
|
Chair: Aksikas, Ilyasse | Qatar University |
Co-Chair: Valean, Honoriu | Technical University of Cluj-Napoca |
|
09:30-09:50, Paper ThA3.1 | Add to My Program |
A Unified Framework for Structural - Temporal Coherence in Graph Neural Networks for Data-Driven Fault Identification |
|
Plakias, Spyridon | Democritus University of Thrace |
Boutalis, Yiannis | Democritus University of Thrace |
Keywords: Fault Diagnosis and Fault Tolerant Control, Neural Networks, Machine Learning
Abstract: In modern industrial systems, accurate fault identification is crucial for the early isolation of broken parts and for further system restoration. Furthermore, data-driven machine learning applications gain popularity because of the increased availability of sensor data and their effectiveness. Graph neural networks, which are neural processes that operate on graph-structured data, are ideal for representing non-Euclidean data captured from multiple sensors. In the current paper, we benefit from the representation ability of graph structures through the presentation of a novel neural graph model that utilizes residual connections and leverages the spatial structure of the graph and the local information of graph nodes. Adaptive structural information and temporal knowledge insight can be integrated by the suggested graph neural framework. The latter is accomplished by feeding the temporal encoding of the graph nodes into a spectral graph neural network for training. Simulation results on the widely used fault identification benchmark of the Tennessee Eastman industrial chemical process verify that the proposed method outperforms competitive machine learning methods and state-of-the-art graph neural models, strengthens graph neural network training, and can be used to accurately identify faults in real industrial scenarios.
|
|
09:50-10:10, Paper ThA3.2 | Add to My Program |
Enhancing Effluent Quality Predictions in Wastewater Treatment with LSTM Neural Network |
|
Voipan, Daniel | Dunărea De Jos University of Galați |
Vasiliev, Iulian | Dunarea De Jos University of Galati |
Condrachi, Larisa | Dunarea De Jos University of Galati |
Voipan, Andreea-Elena | Dunărea De Jos University of Galați |
Barbu, Marian | Dunarea De Jos University of Galati |
Keywords: Neural Networks, Machine Learning
Abstract: This research paper delves into the application of Long Short-Term Memory (LSTM) neural networks within the Benchmark Simulation Model No. 2 (BSM2) to enhance the predictability and efficiency of wastewater treatment processes. The study aims to develop advanced predictive models that can simulate the dynamics of wastewater treatment more accurately and adjust operational strategies dynamically. By integrating LSTM networks, the research enables continuous prediction of Effluent Quality Index (EQI) variables under stochastic and deterministic scenarios, thereby improving the accuracy and efficiency of predicting pollutant levels. The research uses a LSTM model to learn from a comprehensive dataset derived from historical simulations of BSM2, where key parameters such as the oxygen transfer coefficient (KLa) are systematically varied to measure their impact on effluent quality. The LSTM's capability to handle complex, nonlinear data and its adaptability to time series forecasting significantly enhances model performance, offering a robust tool for real-time decisionmaking and process optimization in wastewater treatment facilities. This approach not only improves the accuracy and efficiency of predicting pollutant levels but also supports environmental compliance and operational sustainability, making it a valuable tool for environmental engineers and professionals in the field of wastewater treatment.
|
|
10:10-10:30, Paper ThA3.3 | Add to My Program |
Control Oriented Neural Network Model Learning : L2-Disturbance Attenuation Via an Approximate Input-Output Linearizable Model |
|
Yagoubi, Mohamed | IMT Atlantique (LS2N)/ARMINES |
Hache, Alexandre | IMT Atlantique |
Thieffry, Maxime | IMT Atlantique |
Chevrel, Philippe | IMT Atlantique, LS2N (UMR 6004) |
Keywords: Neural Networks, Nonlinear Systems, System Identification and Modeling
Abstract: This paper explores the utilization of machine learning techniques to develop an approximate input-output linearizable neural network model aimed at improving disturbance attenuation. The incorporation of a learning mechanism allows for the synthesis of control laws that effectively address disturbance attenuation by imposing a specific parameterization and an mathcal{L}_2 gain constraint during the learning process. The constraint is subsequently relaxed into a set of diagonally dominant (DD) matrix constraints. This relaxation leads to a series of linear constraints that can be seamlessly incorporated into the loss criterion. Therefore, a log-sum-exp (LSE) function—a smoothed version of the max function— of these linearized constraints is added to the loss criterion which results in an unconstrained problem amenable to training via back-propagation. The proposed methodology is applied to two variants of nonlinear perturbed pendulum systems. The results emphasize the effectiveness of the model, both on its own and as a framework for developing control laws for disturbance attenuation.
|
|
10:30-10:50, Paper ThA3.4 | Add to My Program |
Enhancing Wastewater Treatment Sensor Fault Detection through Deep Learning |
|
Ghinea, Liliana Maria | University Dunarea De Jos Galati |
Miron, Mihaela | Dunarea De Jos University |
Barbu, Marian | Dunarea De Jos University of Galati |
Keywords: Neural Networks, Fault Diagnosis and Fault Tolerant Control, Nonlinear Systems
Abstract: Efficient operation and maintenance of wastewater treatment plants (WWTPs) are essential for safeguarding public health and the environment. The emergence of mechanical faults within complex systems can lead to disruptions, increased operational costs, and environmental risks. As the world moves towards a digitally connected and sustainable future, the development of Deep Learning (DL) tools for fault detection and isolation (FDI) in wastewater treatment processes is expected to become paramount. Therefore, in this study, we developed two neural models, a Feedforward Neural Network (FFNN) and a Long Short-Term Memory (LSTM), to address the detection of mechanical faults such as bias, stuck, spikes, and precision degradation of the Dissolved Oxygen (DO) sensor. The classification results showed remarkable accuracy performances during testing: for Dataset 1, FFNN achieved 96.56%, while LSTM reached 99.36%; and for Dataset 2, FFNN achieved 99.36%, and LSTM reached 99.57%.
|
|
10:50-11:10, Paper ThA3.5 | Add to My Program |
Colouring Grayscale Images Using Cycle-Generative-Adversarial Networks |
|
Hobinca, Radu-Cosmin | Gheorghe Asachi Technical University of Iasi |
Ferariu, Lavinia | Gheorghe Asachi Technical University of Iasi |
Keywords: Neural Networks, Machine Learning
Abstract: This paper analyses models that can realistically map the grayscale image domain to the color image domain. We thus focus our attention on a CycleGAN (Cycle-consistent Generative Adversarial Network) neural network architecture, which proves to have good results in the area of translation between domains, and implicitly in the application of image coloring. In addition to exploring CycleGAN for image colorization, this study introduces novel techniques aimed at enhancing its performance. We incorporate a color distribution loss term to ensure reliable color mapping, effectively addressing discrepancies in color distribution between domains. Moreover, an in-depth analysis of generator errors is conducted, unveiling critical insights into model limitations. Leveraging this analysis, we propose a compensatory network to generate error images, facilitating more accurate colorization. Through rigorous experimentation across diverse datasets, our approach showcases a remarkable capacity to produce lifelike and coherent colorizations.
|
|
11:10-11:30, Paper ThA3.6 | Add to My Program |
Skin Cancer Diagnosis Using CNN with Attention Mechanism Based on Grad-CAM |
|
Prisacariu, Ana Maria | Gheorghe Asachi Technical University of Iasi |
Ferariu, Lavinia | Gheorghe Asachi Technical University of Iasi |
Keywords: Intelligent Systems, Machine Learning, Neural Networks
Abstract: Skin cancer is one of the most prevalent types of cancer, with approximately 2 to 3 million cases diagnosed globally each year. Typically originating in skin areas exposed to the sun, deviations in the deoxyribonucleic acid (DNA) of skin cells can also precipitate its emergence in other regions. The paper discusses an application for the diagnosis of a diverse spectrum of dermatological conditions, focusing on the design of convolutional neural network (CNN) architectures capable of providing an accurate classification. The input data consists of dermoscopic images, and the associated classes indicate the type of lesions. CNN architectures with single or multiple attention mechanisms are analyzed in this context. Spatial attention mechanisms are applied to different layers of the feature extractor, to highlight local and global features from the most relevant areas of the images. This paper also introduces attention mechanisms that exploit Grad-Classification Activation Maps (Grad-CAMs) resulting from previously trained models, for an effective transfer of knowledge to the attention weights. The experimental demonstration was done for the AlexNet neural network, using images from the Human Against Machine (HAM10000) dataset, corresponding to 5 dermatological conditions. The best accuracy on testing data was obtained for models with single attention mechanisms related to global features. In addition, these models demonstrate a better understanding of dermoscopic images, as illustrated by the mean Grad-CAMs.
|
|
ThA4 Regular session, Constantin Brancusi |
Add to My Program |
Software Methods and Tools |
|
|
Chair: Stamatescu, Iulia | University Politehnica of Bucharest |
Co-Chair: Popescu, Paul-Stefan | University of Craiova |
|
09:30-09:50, Paper ThA4.1 | Add to My Program |
Integrating Logical Dependencies in Software Clustering: A Case Study on Apache Ant |
|
Stana, Adelina Diana | Universitatea Politehnica Timisoara |
Sora, Ioana | University Politehnica Timisoara |
Keywords: Software Methods and Tools, Information Systems Applications
Abstract: Logical dependencies, extracted from co-changes from the versioning system, have multiple applications across numerous fields, including fault detection, software reconstruction, key class identification, among others. This paper will focus on the influence of code co-changes on software clustering for architectural reconstruction. Specifically, we will analyze their impact on the clustering solution of Apache Ant in order to assess whether co-changes usage enhances the quality of the obtained solution.
|
|
09:50-10:10, Paper ThA4.2 | Add to My Program |
CUDA Accelerated Graph Algorithms for Key Class Detection |
|
Chirila, Ciprian-Bogdan | Politehnica University of Timisoara |
Sora, Ioana | University Politehnica Timisoara |
Keywords: Software Methods and Tools, Computational Methods
Abstract: Key classes are deemed as pivotal elements within a software system, serving as focal points for reengineering or documentation endeavors. The identification of these key classes holds significant importance in contemporary practices, with numerous studies dedicated to automating their detection based on representations within class graphs. Research indicates that employing algorithms such as Hyperlink-Introduced Topic Search (HITS) and PageRank (PR) yields optimal precision and recall performance in identifying key classes. However, the runtime execution of these algorithms becomes critical, particularly when operating on graphs with varied weights attributed to class relationships. To address the challenge of runtime execution, we explore parallel implementations of these algorithms utilizing CUDA, invoked from a Java application through JCuda. Specifically, we investigate two approaches: i) employing Java virtual threads and ii) utilizing CUDA threads within the context of the JCuda library. CUDA has fundamentally transformed how we harness GPU acceleration across diverse computational tasks, spanning parallel processing, deep learning, and high-performance computing. Our experiments are conducted on a data set comprising 14 Java projects. Our findings reveal that the hardware parallel CUDA threading model significantly accelerates attribute computation, achieving a runtime reduction of 95% to 97% compared to the virtual threading model.
|
|
10:10-10:30, Paper ThA4.3 | Add to My Program |
Algorithmic Approach for Audio Segmentation |
|
Sarmasanu, Vasile Silviu | Technical University Gheorghe Asachi Iasi |
Butincu, Cristian Nicolae | "Gheorghe Asachi" Technical University of Iași |
Manta, Vasile | Gheorghe Asachi Technical University of Iasi |
Stoleru, Andrei | Gheorghe Asachi Technical University of Iași |
Keywords: Software Methods and Tools
Abstract: This paper presents an algorithmic approach for segmenting news broadcasts produced by television stations. The proposed approach uses advanced signal processing techniques to extract and classify audio segments from Romanian TV news broadcasts. By using derivative of 1-dimensional Gaussian filter and binary classification models, this paper aims to automatically identify and classify news segments into categories such as studio news, field news, commercials and music breaks. The primary objective of this paper is to develop a segmentation method- ology optimized for reduced computational memory and quick processing. By leveraging efficient computational techniques, the news segmentation can be performed in a timely manner without compromising accuracy. This optimization is crucial for real-time or near-real-time applications within TV networks, facilitating tasks such as content monitoring and analysis.
|
|
10:30-10:50, Paper ThA4.4 | Add to My Program |
Problem Solving E-Learning Platform Using Algorithm Visualization and Supervised through Artificial Intelligence |
|
Resceanu, Ionut Cristian | University of Craiova |
Despina, Cosmin | University of Craiova |
Cismaru, Stefan Irinel | University of Craiova |
Petcu, Florina | University of Craiova |
Keywords: Cloud Computing, Computational Methods, Neural Networks
Abstract: As people tend to integrate technology more and more in their lives, one of the most evident aspects which has a great impact on their learning methods are the E-learning platforms which provide interactive means to develop new skills in problem-solving topics. The development of an E-learning platform that simplifies the way in which users engage with the information while simultaneously enhancing their level of comprehension is, without a question, the fastest and most effective strategy. Trend, in matter of learning techniques, has also brought new challenges and opportunities to enhance assimilation of information. Thus, during the process of learning, focus has to be actively maintained through a series of stimuli. Algorithm visualization blended with Artificial Intelligence is the solution that the paper intends to present as a reaction to the actual inclination.
|
|
10:50-11:10, Paper ThA4.5 | Add to My Program |
Improved Perceptual Representation of Isosurfaces from Volume Data Using Curvature-Based Features |
|
Gavrilescu, Marius | Gheorghe Asachi Technical University of Iasi |
Keywords: Computer Graphics, Computer Vision
Abstract: The representation of relevant information from volume data sets is a challenging task due to the high complexity of the structures and spatial features found in such data. The challenge is to represent such structures and features in a way that makes them easy to visually perceive without causing information overload in the resulting images. To this end, we propose a straightforward means of highlighting various regions from isosurfaces found in volume data, such that the visual perception of important surface details is improved. We use an approach based on curvature analysis to determine variations of the isosurface shape, allowing the accentuation of meaningful surface regions. We show that, while the resulting surface accents alone are enough to improve the display of surface details, combining our method with local illumination significantly contributes to a raised level of perception of the surface shape, as well as to the generation of more comprehensive representations of the underlying data. We present our results through illustrative images of medical CT volumes and perform an evaluation using several state-of-the-art no-reference image quality assessment methods. Additionally, our technique does not require precomputation and is easy to incorporate into existing volume rendering engines.
|
|
11:10-11:30, Paper ThA4.6 | Add to My Program |
Performance Evaluation of GPU-Enhanced Jacobi Algorithm Implementations for Optimizing CFD Poisson Partial Differential Equation Solving |
|
Sbîrnă, Sebastian | Technical University of Denmark |
Sbirna, Liana Simona | University of Craiova, Faculty of Mathematics and Natural Scienc |
Keywords: Computational Methods, Engineering Education, Software Methods and Tools
Abstract: The current work presents an in-depth analysis of several optimizations using GPU parallel computing applied to the Jacobi method for solving Poisson partial differential equations in computational fluid dynamics (CFD). We expand on previous CPU-parallelized Jacobi algorithm research, exploring four GPU-optimized Jacobi method variants: single-threaded, multi-threaded, multi-GPU and a norm-based stopping criterion kernel. These implementations are benchmarked against a multi-threaded CPU baseline. Results indicate that, whereas the single-threaded GPU version is slower than the CPU baseline, multi-threaded GPU versions achieve significant speed gains, especially for larger grid sizes. The multi-GPU version doubles memory bandwidth, enhancing performance for extensive computations, despite overhead for smaller matrices. The norm-stopping criterion kernel offers early convergence for small matrices but at a high overhead cost. Profiling confirms a memory-bound bottleneck, suggesting single-precision and optimized memory access as improvements. Ultimately, multi-threaded GPU kernels substantially outperform the CPU baseline for large-scale CFD problems, establishing GPUs as efficient accelerators for the Jacobi algorithm.
|
|
ThP1 Plenary talk, George Enescu |
Add to My Program |
Plenary Session 1 - Frank ALLGÖWER |
|
|
Chair: Valcher, Maria Elena | Universita' Di Padova |
|
12:00-13:00, Paper ThP1.1 | Add to My Program |
Innovations in MPC: The Promise of Model-Based and Data-Driven Methods |
|
Allgower, Frank | University of Stuttgart |
Keywords: Predictive Control
Abstract: Recent years have shown rapid progress of learning-based and data-driven methods, significantly impacting the field of control, including model predictive control (MPC). In addition to numerous methodological and computational advancements, a substantial number of application studies featuring data- and learning-based MPC are currently being published. In this talk, we will compare model-based and data-based MPC to explore which holds more potential for future impact. Highlighting recent developments, we will focus on two different data-based MPC schemes: one based on the Fundamental Lemma of Willems et al., and the other on the data-informativity paradigm. By providing an overview and introduction to these methods, we will discuss their theoretical properties, suitability for nonlinear systems, and demonstrate their advantages and limitations compared to model-based MPC through various application examples. This critical analysis and comparison aim to offer insights and recommendations for future research directions in the evolving domain of MPC.
|
|
ThB1 Regular session, George Enescu |
Add to My Program |
Control Systems and Applications (2) |
|
|
Chair: van Rossum, Felix | Leuphana University Lueneburg |
Co-Chair: Caruntu, Constantin-Florin | Gheorghe Asachi Technical University of Iasi |
|
16:00-16:20, Paper ThB1.1 | Add to My Program |
Set-Based Actuator Fault Diagnosis in a Two-Tank Plant |
|
Culita, Janetta | "Politehnica" University of Bucharest |
Stoican, Florin | Politehnica University of Bucharest |
Olaru, Sorin | CentraleSupélec |
Keywords: Fault Diagnosis and Fault Tolerant Control, Control Applications
Abstract: This paper analyzes a stuck actuator fault scenario in a fluidic plant described by a multivariable system. Based on data-driven analysis with extensive long-term experiments one obtains a nonlinear dynamical model with explicit bounds on noise and fitting errors. A set membership-based method, which considers the residual set bounds, is proposed to diagnose the fault. The accuracy of the set-based FDI method is proven on a simulated fault event.
|
|
16:20-16:40, Paper ThB1.2 | Add to My Program |
Management and Control Strategy of a Hybrid Islanded AC/DC Microgrid |
|
Smouni, Omaima | Mis Laboratory Picardie Jule Vernes University |
Nachidi, Meriem | University of Valladolid |
Rabhi, Abdelhamid | MIS |
Keywords: Fuzzy Systems/Logic, Decentralized/Distributed Control, Control Applications
Abstract: A reliable operation of an islanded Microgrid (MG) depends on the synchronized functioning of renewable and Distributed Energy Resources (DERs) as well as Distributed Energy Storage Systems (DESSs). In this paper, we propose a hierarchical control approach utilizing multi-loop controllers to achieve decentralized power sharing between the various distributed generators and address the v Voltage/Frequency (V /f) restoration issues in an islanded MG. The control strategy employs droop control as the primary layer to ensure efficient power sharing between distributed generators, and decentralized secondary control to adjust the frequency and voltage, restoring them to their reference values. Moreover, for coordinating the power sources into the DC bus, type-2 Fuzzy Logic Control (FLC2) is used, which ensures reliable system operation. Further, an Energy Management Strategy (EMS) based on Multi-Agent Systems (MAS) is designed to enhance the MG functioning. A hybrid islanded MG test system is established using MATLAB/Simulink environment to assess the effectiveness of our design approach. The obtained results validate the efficacy of the suggested control strategy
|
|
16:40-17:00, Paper ThB1.3 | Add to My Program |
Around a Case of Hugens Synchronization |
|
Danciu, Daniela | University of Craiova |
Popescu, Dan | University of Craiova |
Rasvan, Vladimir | University of Craiova |
Keywords: Distributed Parameter Systems, Delay Systems, Nonlinear Systems
Abstract: The paper considers a mechanical system consisting of two second order local oscillators (of the type Li'{e}nard, in particular, of Van der Pol type), united by an elastic rod with distributed parameters having finite length. This model of Huygens synchronization is shown to be in a one-to-one correspondence with a system of coupled delay differential and difference equations. To these systems an energy-like Lyapunov functional is associated. The Lyapunov functional is shown to be non-increasing at least for large perturbations. The systems are thus Levinson dissipative in the metrics induced by the Lyapunov functional itself. An interesting result is that the local additional damping factors induced em{via} radiation dissipation quench the local oscillations but not the overall ones. Notwithstanding, instability for small deviations, which persists, can lead to overall self sustained oscillations generated by Andronov-Hopf bifurcations or to non-periodic Yakubovich type oscillations.
|
|
17:00-17:20, Paper ThB1.4 | Add to My Program |
Dynamic Subspace Model Predictive Control: A Novel Approach for Improved Computational Efficiency |
|
Dogruer, Can Ulas | Hacettepe University |
Keywords: Constrained Control, Predictive Control, Optimization and Optimal Control
Abstract: Model predictive control has emerged as a prominent technique in control engineering due to its ability to handle constraints on both control signals and system states. This capability makes model predictive control a powerful tool, particularly for complex systems with operational limitations. However, a major challenge associated with model predictive control is the curse of dimensionality arising from the constrained optimization problem solved at each time step. This problem becomes computationally expensive as the system dimension increases. This study proposes an accelerated model predictive control algorithm that addresses the curse of dimensionality. We achieve this by solving an equivalent suboptimal model predictive control problem within a reduced dimensional subspace. The subspace is efficiently calculated using singular value decomposition of the Hessian matrix associated with the quadratic cost function. An adaptation law dynamically determines the subspace size, balancing accuracy and computational efficiency of the model predictive control controller.
|
|
17:20-17:40, Paper ThB1.5 | Add to My Program |
Extension of Integrating Process FOPID Tuning Rules with Performance Degradation in Servo/Regulador Modes Approach |
|
Madrigal, Sebastián | Universidad De Costa Rica |
Arrieta, Orlando | Universidad De Costa Rica |
Rojas, Jose David | Universidad De Costa Rica |
Meneses, Montse | Universitat Autonoma De Bacelona |
Vilanova, Ramon | Universitat Autonoma De Barcelona |
Keywords: Control Systems Design, Optimization and Optimal Control
Abstract: This paper presents an extension to a set of tuning rules for Fractional Order PID controllers based on integrating processes modeled by an Integrating Plus Dead-Time transfer function. The extension is proposed since the basic rules do not consider a performance degradation analysis and therefore fail in the implementation for real processes, the Weighted Performance Degradation methodology is applied to find tuning rules that consider a balance in the trade-off between the servo/regulatory modes of the system; after the optimization phase and the adjustment of the controller parameters, a tuning rule for different degrees of priority between the two modes is proposed and validated by means of concrete examples.
|
|
17:40-18:00, Paper ThB1.6 | Add to My Program |
Model Reference Based Robust Tuning Rule for 2-DoF Fractional PI Controllers with Robustness Constraint for SOPDT Process Model |
|
Campos Salas, Daniel | Universidad De Costa Rica |
Madrigal, Sebastián | Universidad De Costa Rica |
Arrieta, Orlando | Universidad De Costa Rica |
Rojas, Jose David | Universidad De Costa Rica |
Meneses, Montse | Universitat Autonoma De Bacelona |
Vilanova, Ramon | Universitat Autonoma De Barcelona |
Keywords: Control Systems Design, Optimization and Optimal Control
Abstract: This paper presents a methodology for the design of a two-degree-of-freedom fractional-order PI controller (2-DoF FOPI) tuning rule for second-order plus dead-time (SOPDT) models. The methodology, which is based on Model Reference Robust Tuning (MoReRT), is intended to provide a robust and reliable approach to controller tuning. The optimization procedure to define the tuning rule is based on matching the response of the proposed control system with target transfer functions defined from the servo and regulatory control modes of the system. This poses a design methodology that differs from that usually proposed for 2-DoF controllers. Furthermore, MoReRT enables the design of controllers with desired robustness levels within the range of typical values, which typically fall between 1.40 and 2.00 for the maximum value of the sensitivity function. To validate this proposed extension to the original methodology for integer controllers, an illustrative example is presented where the improvement in terms of control system performance and robustness can be quantified.
|
|
ThB2 Invited session, Nicola Iorga |
Add to My Program |
Complex Data Processing for Monitoring, Diagnosis, and Control |
|
|
Chair: Popescu, Dan | University POLITEHNICA of Bucharest |
Co-Chair: Ichim, Loretta | Politehnica University of Bucharest |
Organizer: Popescu, Dan | University POLITEHNICA of Bucharest |
Organizer: Ichim, Loretta | Politehnica University of Bucharest |
|
16:00-16:20, Paper ThB2.1 | Add to My Program |
Exploring the Use of MPLS Technologies in Banking Transactions (I) |
|
Stanescu, Cristian | Valahia University of Targoviste |
Predusca, Gabriel | University Valahia of Targoviste |
Angelescu, Nicoleta | Valahia University of Targoviste |
Circiumarescu, Denisa | University Valahia of Targoviste |
Puchianu, Dan Constantin | Valahia University of Targoviste |
Keywords: Communication Systems, Communication Networks
Abstract: This paper utilizes modern transmission technologies as a foundation for discussing traffic management in banking transactions. It begins with a preliminary analysis of existing requirements and challenges. A comprehensive study is then conducted on a network infrastructure based on Multiprotocol Label Switching (MPLS). Through the application of active techniques, this study aims to bring the benefits of MPLS technology closer to implementation for switched optical media. The analysis presented addresses some of the challenges encountered when deploying a system for banking transactions with such characteristics, focusing on aspects like traffic management, quality of service, and network availability.
|
|
16:20-16:40, Paper ThB2.2 | Add to My Program |
The Efficiency of Routing Protocols, OSPF and EIGRP, Using Graphical Network Simulator-3 and Python (I) |
|
Dumitrache, Cezar-Gabriel | Interdisciplinary Doctoral School, University of Pitești |
Ghita, Cosmin-Alexandru | Universitatea Valaahia Din Targoviste |
Predusca, Gabriel | University Valahia of Targoviste |
Circiumarescu, Denisa | University Valahia of Targoviste |
Angelescu, Nicoleta | Valahia University of Targoviste |
Puchianu, Dan Constantin | Valahia University of Targoviste |
Keywords: Communication Networks, Communication Systems
Abstract: The study explores the utilization of Python and Graphical Network Simulator-3 (GNS3) for analyzing the OSPF and EIGRP routing protocols. Python facilitates telecommunications network management by automating network device configuration and troubleshooting. GNS3 integrates support for Python, enabling users to control network devices in a simulated environment. Using Netmiko, an open-source library, the automation of network topology creation and network scenario testing has been demonstrated. This approach provides an efficient method for analyzing the performance of routing protocols, contributing to the development and management of modern networks.
|
|
16:40-17:00, Paper ThB2.3 | Add to My Program |
Cascaded Predictive Speed Control and Load Torque Rejection of PMSM (I) |
|
Costin, Madalin | University "Dunarea De Jos" of Galati |
Lazar, Corneliu | Gheorghe Asachi Technical University of Iasi |
Keywords: Predictive Control, Constrained Control, Control Applications
Abstract: Abstract— In this paper, a cascade control structure is proposed for model predictive control (MPC) of permanent magnet synchronous machine (PMSM). Based on the d-q model of the PMSM and the field-oriented control principle, the cascade structure consists of an inner loop for predictive control of d-q currents and an outer one for predictive control of speed and load torque rejection. To independently control the two currents in the inner loop, a decoupling algorithm was used, followed by MPC controllers that by their ability to manipulate the constraints in the design phase ensure efficient control considering the physical limitations of currents and voltages. Quadratic constraints of electrical signals are approximated by box constraints to reduce computational effort. For speed control and load torque rejection, an MPC controller with a measurable disturbance input is used in the outer loop. Since the load torque is not measurable, an observer was used to estimate it. The proposed predictive control structure for the PMSM was implemented in Matlab-Simulink using MPC controllers from the Model Predictive Control Toolbox. The Simulink control scheme was used for the performance testing against classical PI controllers.
|
|
17:00-17:20, Paper ThB2.4 | Add to My Program |
Tobacco Plant Detection Using a Performant Neural Network (I) |
|
Vicol, Andrei | University POLITEHNICA Bucharest |
Popescu, Dan | University POLITEHNICA of Bucharest |
Ichim, Loretta | Politehnica University of Bucharest |
Keywords: Neural Networks, Machine Learning, Computer Vision
Abstract: Today, plant detection with high accuracy is an important goal in modern agriculture. Achieving this goal makes it possible to monitor individual plant growth and evaluate global production in modern farms. The paper proposes an intelligent system based on the performant YOLOv8 architecture to detect individual plants in the crops. The case study and the experimental results are obtained on a tabaco drop. A small dataset with the augmentation procedures was used to produce reliable object detection and counting. The newly obtained preprocessed dataset is then split into patches and used to obtain a JSON file describing all the labels present in the image. The YOLOv8 model is first trained on a COCO-formatted augmented dataset. The best set of weights is obtained by evaluating the performance metrics over the whole training phase, picking the one that offers the best compromise between performance, training duration, and risk of induced overfitting.
|
|
17:20-17:40, Paper ThB2.5 | Add to My Program |
Model-Free Adaptive Pitch Control for a Nonlinear Aerospace Laboratory Equipment (I) |
|
Baciu, Andrei | "Gheorghe Asachi" Technical University of Iasi |
Lazar, Corneliu | Gheorghe Asachi Technical University of Iasi |
Keywords: Data - Driven Control, Control Applications, Control Systems Design
Abstract: The Model Free Adaptive Control (MFAC) control laws are attractive for applications involving complex plants. This is mainly due to the fact that no process modeling is required, which becomes more and more cumbersome, but only a dynamic linearization through the concept of pseudo partial derivatives (PPD), which it introduces. Using PPD, the control law uses only the input and output data of the process to understand and control it. Systems in the aerospace category are increasingly complex, using a multitude of sensors, information that can be used very easily for applications involving MFAC. The paper proposes a test of an MFAC law conjugated with Sliding Mode Control (SMC), to test this type of law in comparison with the classic MFAC-CFDL variant.
|
|
17:40-18:00, Paper ThB2.6 | Add to My Program |
Cloud Architectural Design for Image-Based Vehicle Positioning in Traffic Management (I) |
|
Beți, Iosif-Alin | Gheorghe Asachi Technical University of Iasi |
Herghelegiu, Paul | Technical University "Gheorghe Asachi" |
Amarandei, Cristian-Mihai | "Gheorghe Asachi" Technical University of Iasi, Faculty of Autom |
Caruntu, Constantin-Florin | Gheorghe Asachi Technical University of Iasi |
Keywords: Cloud Computing, Computer Vision, Automotive Control Systems
Abstract: Vehicle positioning algorithms are essential for improving traffic management and safety by accurately locating vehicles in real-time, and, thus, minimizing congestion and accidents. They also support the development of advanced driver assistance systems and autonomous vehicles, relying on precise positioning data for safe navigation. One of the solutions involves using image processing algorithms, which can have two approaches. One approach is decentralized, in which each vehicle performs its own computing steps and determines its position concerning the other nearby vehicles. The second approach, proposed in this paper, is centralized, where each vehicle sends data to a server that uses cloud computing to process all the data in real-time. As such, vehicles can create a more comprehensive view of the driving conditions in the area by using either of these two approaches, which can help them anticipate potential hazards and make more informed decisions.
|
|
18:00-18:20, Paper ThB2.7 | Add to My Program |
Acquisition and Monitoring of Environmental Parameters from High-Altitude UAV Flight (I) |
|
Dima, Marius | AFT R&D |
Popescu, Dan | University POLITEHNICA of Bucharest |
Ichim, Loretta | Politehnica University of Bucharest |
PÂrvu, Petrisor | POLITEHNICA of Bucharest |
Keywords: Aerospace Systems, Autonomous Systems, Sensor Networks
Abstract: Unmanned aerial vehicles can improve short-term weather forecasting by acquiring information from weather sensors but also from other types of sensors. With this information, there is the possibility of making relevant maps such as solar radiation, pollen, emissions, particles, and others. Another advantage of this acquisition system is the high rate of flights, compared to the classic measurements made with the weather balloon, which is launched twice a day. In addition to the advantages listed above, we can discuss the multiplication factor of the acquired data, these systems can operate in various geographical locations.
|
|
ThB3 Regular session, Mircea Eliade |
Add to My Program |
Aerospace Systems & Industrial Applications |
|
|
Chair: Nitulescu, Mircea | University of Craiova |
Co-Chair: Ifrim, George | Dunarea De Jos University of Galați |
|
16:00-16:20, Paper ThB3.1 | Add to My Program |
Robust Spacecraft Attitude Reference Tracking Using Tube-Based Model Predictive Control |
|
Meinert, Alexander | E: Fs TechHub GmbH |
Baldauf, Niklas | E: Fs TechHub GmbH |
Turnwald, Alen | E: Fs TechHub GmbH |
Keywords: Aerospace Systems, Adaptive and Robust Control, Predictive Control
Abstract: This paper proposes a two-layer robust control framework for three-axis attitude reference tracking while maintaining requirements on the spacecraft’s actuators and motion profile in the presence of uncertainty. The approach builds upon a nonlinear Lyapunov controller in the inner-loop that seeks to stabilize the nonlinear spacecraft equations of motion. The resulting closed-loop system is augmented by an outer-loop Tube-Based Model Predictive Controller (TBMPC) that leverages the benefits of robust constraint satisfaction. To verify the performance of the control framework, a Monte-Carlo simulation with 150 samples subject to varying uncertainty is conducted in MATLAB. Key findings include that TBMPC outperforms a conventional Model Predictive Controller (MPC) and maintains the actuator and performance requirements for any admissible disturbance realization. Overall, the results presented in this work prove online applicability of TBMPC for the specified high-dimensional pointing scenario.
|
|
16:20-16:40, Paper ThB3.2 | Add to My Program |
Four-Stage Cascaded Control Scheme Based on Robust Nonlinear Dynamic Inversion Technique for Quadrotors |
|
Micu, Mihai | National University of Science and Technology |
Lungu, Mihai-Aureliu | University of Craiova |
Chen, Mou | NUAA |
Ebrahimpour, MohammadReza | Tarbiat Modares University |
Keywords: Aerospace Systems, Adaptive and Robust Control, Control Applications
Abstract: This work proposes a four-stage robust adaptive cascaded control architecture for handling the flight of quadrotors with unknown uncertainties via the dynamic inversion. The main targets of the control architecture are the accurate tracking of the reference trajectory and the robustness in terms of parametric uncertainties. The dynamics of quadrotors is innovatively brought to a strict feedback form and then, a robust control scheme is designed by considering four controllers (for position, velocity, attitude, and angular rate), adaptive control laws (to suppress the uncertainties), and fixed-time command filters. The stability of the closed-loop system is proved via the Lyapunov theory, and then, the proposed control architecture is software validated by complex numerical simulations.
|
|
16:40-17:00, Paper ThB3.3 | Add to My Program |
On the Effects of the Earth’s Infrared Radiation Pressure on Orbiting Satellites |
|
Bouleanu, Daniel-Costel | University of Craiova |
Badica, Costin | University of Craiova |
Barbulescu, Lucian-Florentin | University of Craiova |
Popa, Radu Teodoru | University of Craiova |
Popa, Didi Liliana | Universitatea Din Craiova, Facultatea De Automatica, Calculatoar |
Keywords: Aerospace Systems, Software Methods and Tools, Computational Methods
Abstract: The movement of a satellite in Earth’s orbit is influenced by several perturbing forces. Some of those forces have a significant impact on the satellite dynamic, while others have a much smaller influence. The Earth’s infrared radiation pressure generates one such force. The effect of this force on the orbit of a satellite leads to a change of position of several centimeters up to several meters, depending on the altitude and dimension of the object. This variation can be safely ignored for most satellites, as it is several orders of magnitude smaller than others and its effects can be compensated for easily by the on-board engines during periodic maneuvers. However, there are some cases, like uncontrollable CubeSats, or the GNSS satellites used for accurate position determination, where this perturbation must be considered. This paper presents a software component that computes the effect of the infrared radiation pressure perturbation on satellites orbiting the Earth.
|
|
17:00-17:20, Paper ThB3.4 | Add to My Program |
Data Acquisition Application for Monitoring and Recording the Operation of Generator Blocks |
|
Albita, Anca | University of Craiova, Faculty of Automation, Computers and Elec |
Selisteanu, Dan | University of Craiova |
Popa, Bogdan | University of Craiova |
Stinga, Florin | University of Craiova |
Keywords: Industrial Applications, Distributed Systems, Software Methods and Tools
Abstract: The generator blocks with which the thermoelectric plants are equipped are structures that must satisfy specific operating regimes imposed by the working conditions of the national energy system. To supervise and adjust these structures running mode, continuous monitoring, and specific recording of the parameters afferent to their functioning regimes are required. Since the plant organization and the energy system performance may differ, customized implementations for specific electric power architecture may provide the required monitored information in the optimal way. The present work details the development process and the means of use for a customized data acquisition architecture which meets the monitoring and recording requirements specific to generator block operation. This application currently runs in thermoelectric power plants in Romania. The hardware structure is managed through a software suite, implemented as a handy tool for the monitoring and recording process, through features such as record storing and report generation.
|
|
17:20-17:40, Paper ThB3.5 | Add to My Program |
Extending the Operational Lifespan of Key Components in a Cyber-Physical System. a Method and Case Study |
|
Castano, Fernando | Centre for Automation and Robotics (UPM - CSIC) |
Cruz, Yarens J. | Consejo Superior De Investigaciones Científicas |
Haber, Rodolfo | CSIC |
Villalonga, Alberto | Universidad Politecnica De Madrid |
Keywords: Industrial Applications, Machine/Reinforcement Learning
Abstract: Nowadays extending the lifespan of key electronic components in cyber-physical systems is a must from the viewpoint of efficiency and sustainability. Up-to-date, there is a fresh push to use Artificial intelligence (AI) for estimating the remaining useful life (RUL) of electronic components. This paper proposes an automated framework for predicting RUL and classifying device failures. The framework includes a library of AI-based models to perform RUL estimation and to assess the actual condition of critical electronic components, without disconnecting them in a cyber-physical system. The best model is selected using the Q-learning method, which runs in parallel with the set of models. To evaluate the framework, a case study involving electrolytic capacitors of DC/DC power converters is presented. The results corroborate the suitability of the suggested approach in real-data sets and industrial test-bed.
|
|
17:40-18:00, Paper ThB3.6 | Add to My Program |
Low-Cost Air Pollution Monitoring Systems |
|
Petrica, Silvian | UNSTPB, Politehnica Bucuresti |
Fagarasan, Ioana | University POLITEHNICA of Bucharest |
Stamatescu, Iulia | University Politehnica of Bucharest |
Arghira, Nicoleta | University "Politehnica" of Bucharest |
Pietraru, Radu Nicolae | National University of Science and Technology Politehnica Buchar |
Keywords: Industrial Applications, Information Systems Applications, Internet of Things
Abstract: Recently, discussions on air pollution have escalated due to its harmful impact on human health. More and more often we meet a remarkable interest in air quality monitoring, the society we are living in is increasingly approaching methods of acquiring and installing weather and air quality monitoring stations in different locations, whether we are talking about urban or rural areas. While traditional air quality monitoring systems are slightly outdated and they have limitations in providing real-time data at a reasonable cost, technological advances such as the Internet of Things (IoT) offer promising solutions for efficient monitoring in terms of costs, time and energy consumption. Thus, currently smart cities manage to successfully integrate both hardware and software devices to establish complete air quality monitoring systems. This paper proposes an extensive state-of-the-art and a synthesis on the environmental parameters and the air quality index obtained through certain current methods of air quality monitoring. To overcome today’s pollution monitoring systems drawbacks, the air quality index (AQI) was defined according to EU and World Health Organization standards. The AQI was computed based on three different databases in a proposed online application for pollution monitoring and analysis. The calculation method of AQI with existing monitoring systems was compared with current standards AQI for a clear and uniform view of air quality.
|
|
ThB4 Regular session, Constantin Brancusi |
Add to My Program |
Intelligent Systems |
|
|
Chair: Stan, Andrei | Gheorghe Asachi Technical University of Iasi |
Co-Chair: Moldovanu, Simona | Dunarea De Jos University |
|
16:00-16:20, Paper ThB4.1 | Add to My Program |
Travel Planning Optimization System Employing Genetic Algorithms with Multiple Parameters |
|
Rusu, Iuliana-Elena | Gheorghe Asachi Technical University of Iasi |
Alexandrescu, Adrian | Gheorghe Asachi Technical University of Iasi |
Keywords: Information Systems Applications, Distributed Systems, Software Methods and Tools
Abstract: Many people around the world enjoy traveling as a popular leisure activity. The process of trip planning can be time-consuming, requiring travelers to choose between various options and make decisions based on their preferences and interests. This paper addresses the tourist trip design problem by proposing a Travel Planning Optimization System (TPOS), which includes an architecture that employs genetic algorithms to streamline travel planning. The paper's novelty lies in considering numerous travel factors, such as duration of visits, travel time, preferred types of locations, operating hours, destination popularity and the scheduling of places to eat in the daily program, into a single metric. This approach holds great promise in transforming travel planning by providing customized experiences that closely match the unique preferences and interests of each traveler, while using real location-related information in determining an efficient multiple-day itinerary.
|
|
16:20-16:40, Paper ThB4.2 | Add to My Program |
A Method for Computing Social Rank and Relationship Strength Over a Networking Platform |
|
Constantinov, Calin | University of Craiova |
Feraru, Marian-Bogdan | University of Craiova - Faculty of Automation, Computers and Ele |
Popescu, Paul-Stefan | University of Craiova |
Mocanu, Mihai Lucian | University of Craiova |
Dogaru, Dorian | University of Craiova |
Keywords: Intelligent Systems, Communication Networks, Computational Methods
Abstract: Social network analysis is a large and dynamic domain composed of numerous interlinked research directions. Many of them give priority to node computations and consider, as a measure of their importance, node centrality. We present in this paper a more complete and flexible social media content ranking scheme, based on the computation of Social Rank and Relationship Strength between pairs of users in a social network. Capturing the semantics of every edge (targeted by the latter computations) proved to offer a greater chance of producing accurate results. In terms of data modelling, Neo4j, the most popular graph database, was chosen for its analytical capabilities. We demonstrated our approach against a real Facebook dataset. The scheme could be adapted for the specifics of any dataset and can easily be extended to support new interactions or even other platforms. As further validating the model would require access to ground-truth-labelled data that is difficult to obtain, there are certain limitations in our work
|
|
16:40-17:00, Paper ThB4.3 | Add to My Program |
Study of Ontologies and Rule Engine Integration Applied to Sensor Networks |
|
Opranescu, Veronica | National University of Science and Technology POLITEHNICA Buchar |
Anghel, Ana-Magdalena | Universitatea Națională De Știință 5 |
Ionita, Anca Daniela | Universitatea Națională De Știință 5 |
Keywords: Intelligent Systems, Information Systems Applications, Sensor Networks
Abstract: Automatic integration of rule engines with rules from ontologies plays a crucial role in enhancing rule-based recommendation systems. The paper analyzes a selection of technologies in order to define rules in standardized language and tools accessible to experts in the application domain. These rules are subsequently converted into a format executable by a rule engine. An application for the domain of sensor networks is also presented, including rules related to risk and emergency detection.
|
|
17:00-17:20, Paper ThB4.4 | Add to My Program |
Reinforcement Learning for Path Planning and Control of an Autonomous Vehicle with Collision Avoidance |
|
Ibrahim, Kaneewar | University of Rostock |
Husmann, Ricus | University of Rostock |
Weishaupt, Sven | University of Rostock |
Aschemann, Harald | University of Rostock |
Keywords: Machine Learning, Automotive Control Systems, Robotics
Abstract: This paper deals with collision avoidance for an autonomous vehicle using a model-free Reinforcement Learning (RL) algorithm rooted in the actor-critic paradigm. To achieve this objective, the actor network has to generate a collision-free path for an autonomous robot from a start to an end position as well as to follow this desired path accurately. Within this framework, the actor provides a sequence of input signals for the underlying velocity controllers of the robot drives. To accomplish this purpose for a large number of obstacles, it turns out to be essential to sort the algorithm’s input vector regarding the smallest Euclidean distance between an obstacle and the agent as well as to consider the robot’s relative direction. In a first step, the training and evaluation of the agent is performed in a simulated environment. The second step involves the successful experimental validation of the trained actor network on a TurtleBot 3 Burger (TB3B) - a test platform for autonomous robots.
|
|
17:20-17:40, Paper ThB4.5 | Add to My Program |
Learning to Count with Cell Assemblies: A Neuro-Symbolic Approach |
|
Leon, Florin | Gheorghe Asachi Technical University of Iasi |
Keywords: Machine Learning, Neural Networks, Other Topics
Abstract: Recent advancements in artificial intelligence have led to significant results in various domains, including image classification, natural language processing, and mastering complex games. However, current deep neural networks seem to process information differently from humans. Neuro-symbolic methods may offer a promising solution to address this concern. This paper proposes a preliminary cognitive architecture focused on neural cell assemblies, which can combine the adaptability of neural approaches with the explicit reasoning capabilities of symbolic systems. It presents a case study on learning to count, and highlights mechanisms for learning, generalization, and adaptation based on predictive errors.
|
|
17:40-18:00, Paper ThB4.6 | Add to My Program |
Reinforcement Learning for Dynamic Pricing under Competition for Perishable Products |
|
Gadipudi, Srikar Babu | Indian Institute of Technology Madras |
Kalaimani, Rachel Kalpana | Indian Institute of Technology Madras |
Keywords: Reinforcement Learning, Machine/Reinforcement Learning, Agent - Based Systems
Abstract: Dynamic pricing strategies, which adapt prices based on external influences, have gained prominence in various industries. This approach involves frequent adjustments in response to factors like demand fluctuations, market trends, competitor pricing strategies and product perishability. However, accurately predicting demand and modeling price-environment interactions remains a significant challenge. In this paper, we propose the application of reinforcement learning (RL) to develop a dynamic pricing policy. Our focus is on formulating an optimal pricing strategy in a duopoly setting to maximize revenue, outperforming competitor pricing. Unlike previous studies, we eliminate assumptions about demand estimations, characterizing customer behavior through a utility function incorporating price, product quality, ratings and perishability. We employ the Soft Actor-Critic (SAC) algorithm for its sample efficiency and faster convergence, ensuring consistency. This research presents a novel approach to dynamic pricing in competitive markets, offering valuable insights for revenue maximization strategies.
|
| |