Wyniki wyszukiwania - BazTech

1

Determining the accuracy of a digital terrain model based on image data obtained from an unmanned aerial vehicle

Żukowska Marta, Krasuski Kamil, Wierzbicki Damian

Zeszyty Naukowe. Transport / Politechnika Śląska

|

2025

|

z. 129

293--309

EN

This article presents and describes the results of research on determining the accuracy of a Digital Terrain Model (DTM) developed based on image data obtained from an Unmanned Aerial Vehicle (UAV). The Digital Terrain Model was created using image data acquired by an Unmanned Aerial Vehicle, specifically the fixed-wing with electric propulsion, flying at an altitude of 300 meters. The image data were collected during a photogrammetric survey conducted over a mountainous area in 2021. The final elevation values of the Digital Terrain Model were recorded in a GRID format with a spatial resolution of 5 meters. The article also includes a comparison of the DTM elevations with results obtained from the satellite GPS RTK technique. Based on this, an accuracy of elevation determination for different vertical profiles ranged from 0.19 m to 0.24 m was obtained. Moreover, the study also involves the development of a DTM from data acquired by the Unmanned Aerial Vehicle at an altitude of 150 meters. In this case, the accuracy of determining the elevations of the DTM for different vertical profiles ranged from 0.10 m to 0.16 m. The results of the research are very interesting for the application of UAV technology in aerial photogrammetry, particularly in inaccessible areas, especially mountainous regions.

2

Accuracy analysis of Galileo code positioning for UAV

Cybulska-Gac Klaudia, Krasuski Kamil, Ciećko Adam

Zeszyty Naukowe. Transport / Politechnika Śląska

|

2025

|

z. 127

39--55

EN

Official ICAO certification for the Galileo satellite navigation system is currently being implemented for aeronautical applications. Hence, experimental studies are needed to verify the performance of Galileo for kinematic positioning of the user in aviation. The main objective of this work is to present an optimal computational strategy for determining the user's position and the accuracy parameter of Galileo positioning in civil aviation. The paper uses the least squares method and Kalman filtering to calculate the user position. The calculations were performed in two independent Galileo observation processing software, i.e., RTKLIB and Emlid Studio. Galileo navigation and observation data acquired from a DJI Matrice RTK300 unmanned platform was used in the calculations. The Galileo SPP code method algorithm was used to determine the UAV coordinates. The RTKLIB application uses a solution based on the least squares method model to determine user coordinates using the SPP method. The Emlid Studio application, respectively, is based on the Kalman filtering algorithm. On this basis, the UAV positions were determined for the two computational strategies, and the Galileo positioning accuracy was then determined in the form of position errors and RMS errors. The study shows that Emlid Studio software improves Galileo's kinematic positioning accuracy by between 15 and 65% over the results obtained from the RTKLIB solution. The flight tests carried out, the software used, and the computational strategies can be utilized for other global GNSS systems.

3

An Assessment of Long-term Spatial Agnosticism of GNSS Positioning Degradation Risks Due to Ionospheric Conditions

Sikirica Nenad, Hedji Ivan, Mikša Marko, Brčić David, Ciriković Enes, Filjar Renato

TransNav : International Journal on Marine Navigation and Safety of Sea Transportation

|

2025

|

Vol. 19 No. 1

85--89

EN

The Global Navigations Satellite Systems (GNSS) have been evolved into an essential infrastructure of modern civilisation, a public goods, and enabler of rapidly growing number of technology and socio-economic applications. However, GNSS applications often lack fundamental details on GNSS Positioning, Navigation, and Timing (PNT services performance to define and determine their Quality of Service (QoS). The lack of alignment with the core GNSS PNT deprives GNSS applications of assessing the risks of the GNSS PNT utilisation, thus leaving GNSS applications unable to prepare alternatives and mitigate the causes of GNSS PNT performance disruptions. Here we contributed to solution of the problem with the introduction and long-term performance assessment of the risk model of ionospheric-caused GNSS positioning degradation. Called the Probability of Occurrence (PoO), our team defined the risk model of GNSS positioning degradation caused by ionospheric conditions based on the long term observations of occurrences of degraded GNSS positioning performance. In the process of the GNSS risk model validation, the long-term PoO risk models are developed using the annual 2014 stationary GNSS horizontal positioning error observations derived from the GNSS pseudoranges collected at the International GNSS Service (IGS) reference stations situated in polar (Iqaluit, Canada) and sub-equatorial regions (Darwin, Australia). Two GNSS risk models are compared for similarity using statistical methods of Hausdorff distance and Cramér–von Mises statistical test. Research results show that two GNSS risk models are spatially agnostic, since no significant difference in two long-term GNSS risk models is found. The research results supports the conclusion of generality of the PoO GNSS risk model, and its ability to serve GNSS applications developers, operators, and users in determination of the QoS of particular GNSS applications.

4

Znaczenie wartości współczynników DOP w pomiarach RTN

Uznański Andrzej

Przegląd Geodezyjny

|

2025

|

R. 97, nr 7

15--21

PL

Aktualnie najpopularniejszą techniką pomiarową wśród geodetów są satelitarne pomiary RTN, gdyż można je wykonać jednoosobowo, łatwo i bardzo szybko. Jednakże prostota realizacji pomiaru idzie w parze z bardzo zaawansowanymi algorytmami przetwarzania obserwacji, które muszą obliczyć pozycję nawet w ułamku sekundy na podstawie sygnałów, które pokonały ok. 20 tysięcy kilometrów. W praktyce podstawowym problemem jest realna ocena dokładności i wiarygodności wyników pomiarów RTN. Chyba każdy często wykonujący pomiary RTN spotkał się z sytuacją, że mimo zinicjalizowanego odbiornika zdarzały się w opracowaniu wyniki pomiarów RTN obarczone dużymi błędami. Na ekranie kontrolera zawsze wyświetlana jest informacja o inicjalizacji odbiornika satelitarnego, tzw. potocznie „fix” oraz parametr jakości wyniku w metrach i najczęściej także najpopularniejszy akronim pomiarów satelitarnych, czyli bezwymiarowy współczynnik GDOP lub PDOP. Na podstawie wartości tych parametrów geodeta musi podjąć decyzję o akceptacji wyniku pomiaru RTN lub jego odrzuceniu i powtórzeniu pomiaru. Celem pracy była praktyczna weryfikacja, czy istnieje korelacja między wartościami współczynników DOP a dokładnością współrzędnych wyznaczanych z pomiarów RTN? Jak istotne znaczenie ma wartość współczynnika DOP w trakcie pomiaru RTN na dokładność jego wyniku? Dysponując bazą testową w publikacji poruszono także aspekt dokładności wyników pomiarów RTN.

EN

Currently, the most popular measurement technique among surveyors is satellite RTN measurements, as they can be carried out by a single person, easily and very quickly. The simplicity of measurement implementation, however, is accompanied by highly advanced observation processing algorithms, which must calculate the position even in a fraction of a second based on signals that have traveled approximately 20,000 kilometers. In practice, the main problem is the real assessment of the accuracy and reliability of RTN measurement results. Almost everyone who frequently performs RTN measurements has encountered situations where, despite initializing the satellite receiver, the RTN measurement results were burdened with significant errors during processing. The controller screen always displays information about the initialization of the satellite receiver, commonly known as "fix," as well as the quality parameter of the result in meters and most often the most popular acronym for satellite measurements, the dimensionless GDOP or PDOP coefficient. Based on the values of these parameters, the surveyor must decide whether to accept the RTN measurement result or reject it and repeat the measurement. The aim of the study was to practically verify whether there is a correlation between DOP coefficient values and the accuracy of coordinates determined from RTN measurements. How significant is the DOP coefficient value during RTN measurement for the accuracy of its result? Using a test database, the publication also addressed the aspect of the accuracy of RTN measurement results.

5

Learning rate tuning as a software optimization technique for efficient deep learning

Belhadri Abdelkader, Benchennane Ibtissem

Przegląd Elektrotechniczny

|

2025

|

R. 101, nr 7

181--190

EN

This article examines software level optimizations for accelerating the convergence of deep learning training by dynami cally adapting the learning rate hyperparameter. First, the underlying motivations for reducing training time are presented. Next, a comprehensive overview of learning-rate scheduling methods is presented. Finally, an empirical evaluation on the CIFAR-10 data set-employing a ResNet-18 architecture-demonstrates that these strategies substantially improve training efficiency without com promising accuracy or increasing energy usage.

PL

W niniejszej pracy przebadano optymalizację poziomu oprogramowania w celu przyspieszenia konwergencji szko lenia w zakresie głębokiego uczenia poprzez dynamiczną adaptację hiper-parametru szybkości głębokiego uczenia. Po pierwsze, przedstawiono podstawowe motywacje zredukowania czasu szkolenia. Następnie, dokonano rozległego przeglądu metod harmonogramu szybkości uczenia. Empiryczna ocena zbioru danych CIFAR-10, stosująca architekturę ResNet-18, pokazuje, że omawiane strategie w istotny sposób poprawiają skuteczność szkolenia bez potrzeby kompromisu w zakresie dokładności lub zwiększenia nakładu zużywanej energii.

6

Inter-turn short circuit fault detection in induction motor using XGBoost, KNN and Random Forest

Chahmi Abdelghani, Benkaihoul Said, Naas Toufik Tayeb, Yildirim Özüpak

Przegląd Elektrotechniczny

|

2025

|

R. 101, nr 2

248--253

EN

In this research endeavor, the focus was directed towards investigating a specific fault occurrence within an induction motor, namely an inter-turn short circuit (ITSC), intentionally induced within phase A of the motor. The employed dataset encompassed both correct operational states and instances afflicted with the aforementioned fault, with parameters such as current flows and torque outputs meticulously recorded and analyzed. When employing a methodology rooted in machine learning, a suite of algorithms was applied to discern and identify the presence of the fault. From among the array of algorithms utilized, the notable contenders included Random Forest (RF), k-nearest neighbors (KNN), and Extreme Gradient Boosting (XGBoost), each meticulously trained and tested on the dataset to gauge their efficacy in fault detection. The outcomes obtained in the mentioned study unequivocally demonstrate the superiority of the Random Forest algorithm in terms of accuracy assessment, boasting a remarkable accuracy rate of 99.7%. In the stark contrast, both KNN and XGBoost algorithms exhibited comparatively lower accuracy rates, standing at 96.6% and 96.5%, respectively.

PL

W tym przedsięwzięciu badawczym skupiono się na badaniu konkretnego wystąpienia usterki w silniku indukcyjnym, a mianowicie zwarcia międzyzwojowego (ITSC), celowo indukowanego w fazie A silnika. Zastosowany zbiór danych obejmował zarówno prawidłowe stany operacyjne, jak i przypadki dotknięte wyżej wymienionymi usterkami, przy czym parametry takie jak przepływy prądu i wyjściowy moment obrotowy były skrupulatnie rejestrowane i analizowane. Stosując metodologię opartą na uczeniu maszynowym, zastosowano zestaw algorytmów w celu rozpoznania i zidentyfikowania obecności usterki. Wśród szeregu wykorzystywanych algorytmów godnymi uwagi konkurentami byli Random Forest (RF), k-najbliżsi sąsiedzi (KNN) i Extreme Gradient Boosting (XGBoost), każdy skrupulatnie przeszkolony i przetestowany na zbiorze danych w celu oceny ich skuteczności w wykrywaniu usterek. Wyniki uzyskane w tym badaniu jednoznacznie wskazują na wyższość algorytmu Random Forest pod względem oceny dokładności, który może pochwalić się niezwykłym współczynnikiem dokładności wynoszącym 99,7%. Dla kontrastu, zarówno algorytmy KNN, jak i XGBoost wykazywały stosunkowo niższe wskaźniki dokładności, wynoszące odpowiednio 96,6% i 96,5%.

7

Numerical Analysis of the Discrete, Fractional Order PID Controller using FOBD and CFE Approximations

Oprzędkiewicz Krzysztof

Pomiary Automatyka Robotyka

|

2025

|

R. 29, nr 3

5--11

EN

This paper presents the numerical analysis of the discrete, approximated Fractional Order PID Controller (FOPID). The fractional parts of the controller are approximated with the use of the most known methods: Fractional Order Backward Difference (FOBD) and Continuous Fraction Expansion (CFE). CFE is simpler and faster than the FOBD method, but its accuracy is not always satisfying. For both approximations optimum sample time was found by minimizing of the cost function Integral Absolute Error (IAE). Additionally, to optimize of CFE its parameter a was applied. Results of numerical tests show that the FOPID using FOBD is more accurate in the sense of IAE cost function for FOPI and FOPID controllers, but CFE is more accurate for FOPD controller. Next, the FOBD requires to use of smaller sample time to obtain of good accuracy than CFE. This allows to conclude that FOPD controller using CFE can be applied in time critical applications at bounded platforms, for example in robotics or numerical control.

PL

W pracy zaprezentowano analizę numeryczną dyskretnego regulatora PID niecałkowitego rzędu, w którym akcje: całkująca i różniczkująca są aproksymowane z użyciem dwóch typowych aproksymacji dyskretnych: FOBD i CFE. CFE jest szybsza i prostsza, natomiast nie zawsze zapewnia wystarczającą dokładność. Dla obu badanych aproksymacji wyznaczono okres próbkowania zapewniający ich najlepszą dokładność w sensie funkcji kosztu IAE. W przypadku aproksymacji CFE w optymalizacji wykorzystano dodatkowo współczynnik a. Wyniki testów numerycznych wskazują, że zastosowanie aproksymacji FOBD zapewnia lepszą dokładność dla regulatorów FOPID i FOPI, natomiast dla regulatora FOPD lepszą opcją jest zastosowanie CFE. Regulator FOBD dla zapewnienia dobrej dokładności wymaga stosowania krótszego okresu próbkowania, niż CFE. Podsumowując, w krytycznych czasowo aplikacjach pracujących na sprzęcie o ograniczonej mocy obliczeniowej (np. robotyka, sterowanie numeryczne lub urządzenia IoT) można rekomendować zastosowanie regulatora FOPD wykorzystującego aproksymację CFE

8

Impact of the Datum Point Location on the Benchmark Standard Deviations

Mickrenska Christina, Cvetkov Vasil

Inżynieria Mineralna

|

2025

|

Vol. 3, no. 2

art. no. 12

EN

Precise leveling is the basic method for estimating sea level changes, geoid/quasi-geoid determination, and recent vertical movements of the Earth’s crust for more than 150 years. Despite the long-term usage of the method, there are some issues about the accuracy of leveling networks. It is supposed that the propagation of leveling errors is proportional to the square root of the leveled distance. However, there is a lack of correspondence between the leveling accuracy given by the least squares mean error per 1 km leveling distance and the final standard deviations of the adjusted benchmark heights as regards their remoteness from the datum point. To investigate the sources of this accuracy disagreement, we adjusted the precise leveling network of the Bulgaria – III leveling cycle 33 times. A different nodal benchmark was chosen as a datum point in each adjustment. All adjustments produced the same adjusted benchmark elevations and mean adjustment error per 1 km leveling distance, equal to 1.21 mm.km-0.5. However, the sums of the standard errors of the adjusted heights of the nodal benchmarks, and as a result, the average of these standard errors varied between adjustments. The minimum sum of the benchmark standard errors was obtained when the datum point was set in the Knezha benchmark, located almost in the center of the network. The average standard error of the benchmark in this adjustment was estimated to be 10.89 mm. The official datum point in the Third Leveling of Bulgaria is the fundamental benchmark in Varna, located in the network periphery. The average standard error of the benchmark in an adjustment with a datum point in Varna was estimated to be 13.68 mm. Comparison between the means of the benchmark standard error samples based on the datum point in Knezha and Varna, performed by t-test under the assumption of equal sample variances, shows that the mean of the standard error sample based on the datum point in Knezha is significantly less than in Varna. The conclusion is valid at a 99% confidence level. In addition, a better correspondence between the leveling accuracy given by the least squares mean error per 1 km leveling distance and the final standard deviations of benchmarks was found in the case of the Knezha datum point adjustment.

9

Comparative Analysis of Classical Geodetic Methods and Terrestrial Laser Scanning for Tilt Determination of a High Metal Lattice Structure

Ilieva Emiliya, Angelov Antonio, Borisova Petya

Inżynieria Mineralna

|

2025

|

Vol. 2, no. 2

art. no. 28

EN

This study aims to evaluate the applicability and accuracy of terrestrial laser scanning (TLS) compared to classical geodetic methods for determining the tilt of a high metal lattice structure. The structure under investigation is a 65.5-meter-high lighting mast characterized by complex spatial geometry and an open lattice design, which presents significant challenges for deformation monitoring using conventional surveying techniques. Deviations from the vertical axis were assessed through a series of horizontal cross-sections, analyzed using both total station measurements and TLS point cloud data. Classical geodetic measurements involved angular intersections and trigonometric leveling, performed within a local geodetic control network. TLS data were acquired using a Trimble TX6 scanner and processed through a registration process and georeferenced using target spheres. To evaluate the consistency and accuracy of both methods, graphical comparisons and analyses of the measured values were performed. The results demonstrate a high degree of agreement between the two approaches, with deviations remaining within a few centimeters. These results confirm the potential of TLS as a reliable and efficient method for monitoring structural deformations in high-rise engineering structures with complex geometries.

10

Adjustment of a Precise Leveling Network with Jackknife Selected Weights

Cvetkov Vasil, Mickrenska Christina

Inżynieria Mineralna

|

2025

|

Vol. 3, no. 2

art. no. 10

EN

Many publications indicate no agreement between oceanographic, GNSS, and precise leveling data. The disagreement in the results, produced by different methods, is usually explained by the “mystic” systematic errors in the spirit leveling. All these studies processed the leveling data under the assumption of accumulation of leveling errors as a function of leveling lines. However, according to some recent findings, the length of leveling lines is not the most important factor for leveling uncertainty. Thus, it is likely that the classical weights used in the adjustment of precise leveling networks, which are inverse-proportional to the length of the leveling line, are not the best choice. To avoid any subjectivity in the choice of weights, we used a non-parametric approach for constructing our weights in the current research. We adjusted the Third Finnish Levelling Network 80 times using the Jackknife resampling. We skipped a different line in different adjustments and assessed the network accuracy. Based on these results, we formed our weights for each line in the network as a function of produced accuracy. If a skipped line leads to higher accuracy, we gave it a greater weight in the final adjustment. The comparison between the standard deviations of the adjusted benchmark heights produced by the proposed approach and an adjustment of the network with classical weights shows the supremacy of the non-parametric weights. The hypothesis of the equality of the means of the standard deviation samples, formed by both approaches, is rejected at a 99% confidence level by two independent tests – the t-test Paired Two Samples for Means and the non-parametric Wilcoxon Signed-rank test. All those facts mean that the systematic errors in the spirit leveling are not only in the measured elevations but also in the following observation data processing. The proposed weighting approach can be successfully applied to many geoscience data processing, e.g., terrestrial gravimetric networks, networks for recent Earth’s crust movements, sea level changes, etc.

11

Satellite-Based Urban Heat Island Study: A Prisma-Based Systematic Literature Review

Darmawan Soni, Hernawati Rika, Rahmani Shafa

Geomatics and Environmental Engineering

|

2025

|

Vol. 19, no. 6

5--40

EN

Over the years, urban heat island (UHI) has emerged as a significant contributor to global warming, thereby necessitating considerable attention. Currently, satellite technology is a basic tool for the future – particularly, for its effective and efficient urban analysis. Thus, this study aims to assess the progress of existing satellite-based UHI studies by reviewing scientific publications that were released between 1972 and early 2024. Moreover, we observed that 1991 was a pivotal year, marking the integration of satellite technologies into the development of UHI monitoring and identification systems based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, this review methodology examines the UHI phenomenon by focusing on its characteristics based on sensors, algorithms, and accuracy. The results of the systematic review revealed that Landsat and MODIS were the most-deployed sensors for UHI identification and monitoring, while the land surface temperature (LST) indicator and normalized difference vegetation index (NDVI) were the most-deployed algorithms. Regarding accuracy, the integration of satellite sensors and algorithms into UHI studies provides a promising range of accuracies. The review found that the future of satellite-based UHI monitoring is promising, with technological advancements driving the development of effective techniques such as data fusion, gap filling, machine learning (ML), and deep learning. Additionally, Google Earth Engine (GEE) is a cloud-based platform for performing large-scale geospatial analyses, which facilitates the assessments of local, regional, and global-scale UHIs. Finally, the other review findings for future directions indicated that future satellite-based UHI studies will prioritize six crucial points: enhancing data resolution, integrating satellite data with ground-based sensors, artificial intelligence, and ML, climate change modeling, and a global study of UHIs and their impacts.

12

DermoAI-CNN: Leveraging GANs, Mask R-CNN, and Attention Mechanisms for Enhanced Skin Disease Analysis

Prasad Rajesh, Saxena Kushagr, Maurya Ayush, Prajapati Yogendra Narayan

Foundations of Computing and Decision Sciences

|

2025

|

Vol. 50, No. 4

537--569

EN

The study aims to develop an effective and efficient deep learning model for detecting skin diseases, as skin diseases rank as the world's number one health problem. Besides, cancers and dermatological anomalies should be diagnosed at an early stage, so that subsequent treatment can be efficient and complication-free. The existing methods of diagnosis are associated with lower precision and, in most cases, are inefficient, which can be attributed to the lack of effective data augmentation, segmentation techniques, and improved feature extraction. In this paper, a general framework is introduced that uses Generative Adversarial Networks for data augmentation, Mask R-CNN for precise segmentation, and a tailored multilayer Convolutional Neural Network with an attention mechanism incorporated into it to classify 23 skin disease classes using 25,250 images, among them 5,750 generated by GAN, to balance underrepresented classes. The accuracy attained was 97.30%, which was much better than that reported in earlier studies, which ranged from 85 to 92. The metrics, including an accuracy of 95.65%, a recall of 97.09%, and an F1-score of 96.98%, were used to assess the system's performance in classifying invisible dermatological images. The scalable system provides explanations that support real-time diagnosis, preventing delays and acute health costs. The findings fully fulfil the capabilities of deep learning in dermatology, as the initial diagnosis of the skin disease is accurate, accessible and efficient.

13

A new approach for diabetes risk detection using quadratic interpolation flower pollination neural network

Polly Yulianto Triwahyuadi, Fanggidae Adriana, Ledoh Juan Rizky Mannuel, Pah Clarissa Elfira Amos, Djahi Bertha S., Tupen Kisan Emiliano Rape

Applied Computer Science

|

2025

|

Vol. 21, no. 2

63--81

EN

This study aims to evaluate and compare five algorithms in diabetes detection, namely Flower Pollination Neural Network (FPNN), Particle Swarm Optimization Neural Network (PSONN), Bat Artificial Neural Network (BANN), Stochastic Gradient Descent (SGD), and Quadratic Interpolation Flower Pollination Neural Network (QIFPNN). These algorithms were tested on a diabetes risk dataset divided into training, validation, and testing subsets. The evaluation was based on three main aspects: accuracy, F1 score, and training time. Experimental results showed that QIFPNN outperformed others with an average accuracy of 97.90% and an F1 score of 98.30%, although it required the longest training time (4107.89 seconds). FPNN and BANN achieved competitive accuracy (97.34% and 97.43%) and F1 scores (97.84% and 97.91%), while SGD offered a favorable trade-off with accuracy of 96.87%, F1 score of 97.42%, and the shortest training time (584.50 seconds). PSONN performed less well with an average accuracy of 89.26% and an F1 score of 91.45%. These results indicate that QIFPNN can be relied upon as an effective diabetes risk detection model with superior predictive performance. Although the training time of QIFPNN is longer due to its sophisticated optimization process, this is only a concern during model development, as the final trained model can be efficiently used for real-time prediction in practical applications.

14

Evaluation of the force meter accuracy in the engineering industry applications

Tomczyk Krzysztof, Kowalczyk Małgorzata, Gębarowski Tomasz, Felhő Csaba

Advances in Science and Technology. Research Journal

|

2025

|

Vol. 19, no. 10

65--79

EN

This paper presents an assessment of the force meter (dynamometer) dynamic accuracy applied in the engineering industry. This assessment was made by determining the dynamic error based on the absolute error criterion. The mathematical basis for determining the dynamometers error was presented, and the corresponding procedures were developed. The values of dynamometer parameters and associated uncertainties were determined, then they were compared with the values calculated in the corresponding datasheet and finally the relationship between the dynamic absolute error and the dynamometer parameters was presented. The solutions presented allow the easy and quick determination of the accuracy of the dynamometers, and first of all significantly an increase the reliability and safety of engineering systems which use these type of measuring devices. The upper bound of the dynamic error developed in the paper may constitute an additional comparative criterion in assessing the accuracy of dynamometers produced by different manufacturers. New achievement of this paper is the presentation of the relationship between the dynamic error expressed by using the absolute error criterion and the values of the parameters of the mathematical model of the dynamometer in terms of applications in the engineering industry. These relationships are presented using the corresponding mathematical functions as well as graphically using 3D graphs. The calculation results were obtained using MathCad 5.0 and MATLAB R2024.

15

Algorithms to improve unmanned aerial vehicle positioning accuracy using European Geostationary Navigation Overlay Service and System for Differential Corrections and Monitoring ionospheric corrections

Krasuski Kamil, Bakuła Mieczysław, Gołda Paweł, Ciećko Adam, Grunwald Grzegorz, Mrozik Magda, Kozuba Jarosław

Advances in Science and Technology. Research Journal

|

2025

|

Vol. 19, no. 1

284--300

EN

This paper presents a modified algorithm for determining the positioning accuracy of a UAV based on a joint GPS/EGNOS+GPS/SDCM (Global Positioning System/European Geostationary Navigation Overlay Service+Global Positioning System/ System for Differential Corrections and Monitoring) solution. Firstly, the average weighted model for determining the position of the UAV (Unmanned Aerial Vehicle) was developed. The algorithm takes into account the coordinates from the individual GPS/EGNOS and GPS/SDCM solution as well as correction coefficients that are a function of the inverse of the ionospheric VTEC (Vertical TEC) delay. Next the accuracy term was estimated in the form of the position errors and RMS (Root Mean Square) errors. Finally the Kalman filter algorithm was used for improved the position errors and RMS errors. The developed algorithm is concerned with determining the positioning accuracy of the UAV for BLh (B-Latitude, L-Longitude, h-ellipsoidal height) ellipsoidal coordinates. The algorithm was tested on kinematic GPS/SBAS (Global Positioning System/Satellite Based Augmentation System) data recorded by a GNSS (Global Navigation Satellite System) receiver placed on a DJI Matrice 300RTK type unmanned platform. As part of the research test, two flights of the UAV were performed on 16 March 2022 in Olsztyn. In the first flight, the proposed algorithm enabled an increase in UAV positioning accuracy from 4% to 57% after Kalman filter process. In the second flight, on the other hand, UAV positioning accuracy was increased from 6% to 42%. The developed algorithm enabled an increase in UAV positioning accuracy and was successfully tested in two independent flight experiments. Ultimately, further research is planned to modify the algorithm with other correction coefficients.

16

Accuracy analysis of UAV coordinates using EGNOS and SDCM augmentation systems

Krasuski Kamil, Wierzbicki Damian, Kirschenstein Małgorzata, Żukowska Marta, Gołda Paweł

Zeszyty Naukowe. Transport / Politechnika Śląska

|

2024

|

z. 123

75--100

EN

SBAS systems are applied in precise positioning of UAV. The paper presents the results of studies on the improvement of UAV positioning with the use of the EGNOS+SDCM solutions. In particular, the article focuses on the application of the model of totaling the SBAS positioning accuracy to improve the accuracy of determining the coordinates of UAVs during the realisation of a test flight. The developed algorithm takes into account the position errors determined from the EGNOS and SDCM solutions. as well as the linear coefficients that are used in the linear combination model. The research was based on data from GPS observations and SBAS corrections from the AsteRx-m2 UAS receiver installed on a Tailsitter platform. The tests were conducted in September 2020 in northern Poland. The application of the proposed algorithm that sums up the positioning accuracy of EGNOS and SDCM allowed for the improvement of the accuracy of determining the position of the UAV by 82-87% in comparison to the application of either only EGNOS or SDCM. Apart from that, another important result of the application of the proposed algorithm was the reduction of outlier positioning errors that reduced the accuracy of the positioning of UAV when a single SBAS solution (EGNOS or SDCM) was used. The study also presents the effectiveness of the proposed algorithm in terms of calculating the accuracy of EGNOS+SDCM positioning for the weighted average model. The developed algorithm may be used in research conducted on other SBAS supporting systems.

17

Proposals for the implementation of a decision support system for air defence fire control based on fuzzy networks of targets

Volkov Andrii, Stadnichenko Volodymyr, Yaroshchuk Vitalii, Halkin Yurii, Tokar Oleksandr

Systemy Logistyczne Wojsk

|

2024

|

Vol. 61, no. 2

211--228

EN

The purpose of the research was to improve the control of air defence firepower using fuzzy networks of target installations, enhancing the efficiency and accuracy of defensive actions. The research niche of this article is the optimization of decision support systems in air defence through the application of fuzzy logic to improve real-time threat assessment and response accuracy. The study hypothesized that the integration of fuzzy networks into air defence fire control would lead to improved decision-making accuracy and reduced response time under conditions of uncertainty. The methodology involved data collection using radar, acoustic, and infrared sensors; modelling of fuzzy systems with specialized software; the development of fuzzy rules for threat assessment; and the simulation of real combat conditions to evaluate system effectiveness and its integration with existing detection and tracking equipment. The results demonstrated that the proposed decision support system significantly enhances threat assessment accuracy, reduces reaction time, and improves overall air defence effectiveness. Simulation tests confirmed a notable increase in the speed and precision of defensive measures, highlighting the adaptability of the system to dynamic combat conditions. Furthermore, the integration of fuzzy networks with existing detection and tracking technologies facilitated rapid data processing and optimized firepower management, leading to cost reductions. The study contributes to the advancement of decision support methodologies in air defence by introducing an innovative approach based on fuzzy logic, which enhances the accuracy and efficiency of decision-making under conditions of operational uncertainty. Future research should focus on validating the system’s effectiveness in real-world deployments to further refine its performance.

PL

Celem badania było usprawnienie kontroli siły ognia obrony powietrznej poprzez zastosowanie rozmytych sieci instalacji celów, co miało na celu zwiększenie efektywności i precyzji działań obronnych. Niszową problematyką poruszaną w artykule jest optymalizacja systemów wspomagania decyzji w obronie powietrznej, poprzez zastosowanie logiki rozmytej w celu poprawy oceny zagrożeń w czasie rzeczywistym i precyzji reakcji. W badaniu postawiono hipotezę, iż integracja sieci rozmytych z systemami kierowania ogniem obrony powietrznej prowadzi do zwiększenia dokładności podejmowania decyzji oraz skrócenia czasu reakcji w warunkach niepewności. Część badawcza obejmowała zbieranie danych za pomocą radarów, czujników akustycznych i podczerwieni; modelowanie systemów rozmytych przy użyciu specjalistycznego oprogramowania; opracowanie reguł rozmytych do oceny zagrożeń; oraz symulację rzeczywistych warunków bojowych w celu oceny skuteczności systemu oraz jego integracji z istniejącym sprzętem wykrywającym i śledzącym cele. Wyniki badań wykazały, iż zaproponowany system wspomagania decyzji znacząco zwiększa dokładność oceny zagrożeń, skraca czas reakcji oraz poprawia ogólną skuteczność obrony powietrznej. Testy symulacyjne potwierdziły znaczący wzrost szybkości i precyzji działań obronnych, podkreślając zdolność systemu do adaptacji do dynamicznych warunków bojowych. Ponadto integracja sieci rozmytych z istniejącymi technologiami wykrywania i śledzenia celów umożliwiła szybsze przetwarzanie danych i optymalizację zarządzania siłą ognia, co przyczyniło się do redukcji kosztów. Badanie wnosi wkład w rozwój metodologii wspomagania decyzji w obronie powietrznej poprzez wprowadzenie innowacyjnego podejścia opartego na logice rozmytej, które zwiększa dokładność i efektywność podejmowania decyzji w warunkach operacyjnej niepewności. Przyszłe badania powinny skupić się na walidacji skuteczności systemu w rzeczywistych warunkach operacyjnych w celu dalszego doskonalenia jego działania.

18

Investigation of the accuracy of BeiDou, QZSS and QZSS/BeiDou satellites configuration for short, medium and long baselines in the Asia-Pacific regions

Pirti Atinc

Reports on Geodesy and Geoinformatics

|

2024

|

Vol. 118

17--21

EN

The field of satellite navigation has seen significant advancements due to the fast development of multi-constellation Global Navigation Satellite Systems (GNSS). Around 150 satellites will be in service when all six systems – GPS, GLONASS, Galileo, BeiDou, QZSS, and NAVIC – are launched by 2030, offering both enormous potential and advantages for research and engineering applications. This study used an experiment on the accuracy, particularly for short, medium, long baselines (Wide Lane ambiguity solution) of the BeiDou, QZSS and QZSS/BeiDou combinations. It showed that with the integration of BeiDou/QZSS static measurements in the study region millimetre-centimetre accuracy for short, medium, and long baselines can be attained. Based on the results of this study, it can be concluded that the 1st (QZSS/BeiDou), 2nd (BeiDou), and 3rd (QZSS) strategies feature different horizontal accuracies for all categories. The obtained results with different satellite configurations for the Fixed-Wide-Lane integer ambiguity solution are compared with each other. Accuracy at the short baseline (BeiDou, QZSS, and BeiDou/QZSS satellites) was obtained in the range of 0.5–0.7 cm. For the medium baseline, it was computed around 1.8–82 cm. For the long baseline, the accuracy was 5.6–13.3 cm.

19

Ocena jakości obserwacji skanera laserowego velodyne w oparciu o dane symulowane

Pałęcki Adam, Jóźków Grzegorz

Przegląd Geodezyjny

|

2024

|

R. 96, nr 6

32--35

PL

Tanie skanery z wieloma wiązkami laserowymi takie jak Velodyne, Ouster, Hesai często wykorzystywane są do budowy niedrogich systemów skaningu kinematycznego, w tym systemów plecakowych i bezzałogowych. Niski koszt skutkuje mniejszą jakością pozyskiwanych danych, a parametry dokładnościowe podawane przez producentów często odbiegają od rzeczywistych. Z tego powodu problem oceny dokładności danych pozyskanych za pomocą takich skanerów jest ciągle podnoszony przez naukowców. Metody przez nich stosowane mają na celu ocenę dokładności położenia punktów skaningu i opierają się głownie na punktach i powierzchniach referencyjnych. Należy jednak zaznaczyć, że na dokładność położenia tych punktów wpływ mają różne czynniki, w tym te wynikające z błędów instrumentalnych, wynikające z charakteru mierzonego obiektu, a także danych z innych sensorów (np. dane o trajektorii stosowane w skaningu mobilnym). W tym artykule proponujemy metodę, która pozwala na ocenę jakości obserwacji (odległości i kątów), których błędy wynikają głównie z pierwszego z wymienionych czynników, czyli instrumentu. Metoda ta bazuje na porównaniu obserwacji rzeczywistych z teoretycznymi powstającymi poprzez symulację. Do symulacji rzeczywistych obserwacji stosowany jest wirtualny skaner Velodyne, który umieszczany jest w takiej samej pozycji i orientacji jak rzeczywisty. Obserwacje teoretyczne dla skanera wirtualnego tworzone są w oparciu o znany mechanizm działania skanera oraz dokładną i bardzo gęstą chmurę punktów naziemnego skaningu laserowego. Wykonane dla skanera Velodyne HDL-32E eksperymenty wykazały, że dokładność pomiaru odległości jest porównywalna z podawaną przez producenta, jednak inna dla różnych diod laserowych, a dokładność pomiaru kąta poziomego wynosi około 0,04°. Ponadto wykazano, że częstotliwość wirownia skanera, od której zależy wartość kąta poziomego jest różna od wartości nominalnej i nie jest stała w trakcie całego obrotu. Opracowana metoda symulacji obserwacji może być w przyszłości wykorzystana do kalibracji podobnych skanerów tego typu.

EN

Inexpensive scanners with multiple laser beams such as Velodyne, Ouster, Hesai are often used to build low-cost kinematic scanning systems, including backpack and unmanned systems. Low costs result in lower quality of the acquired data. In addition, the accuracy parameters provided by manufacturers are often different from the actual ones. For this reason, the problem of assessing the accuracy of data obtained using such scanners is investigated by scientists. The methods used for this purpose aim at assessing the accuracy of the position of scanning points and use mainly reference points and surfaces. However, that the accuracy of the location of these points is influenced by various factors, including those resulting from instrumental errors, from the nature of the measured object, as well as data from other sensors (e.g. trajectory data used in mobile scanning). In this article, we propose a method that allows for the assessment of the quality of observations (distances and angles) which errors result mainly from the first of the mentioned factors, i.e. the instrument. Proposed method bases on the comparison of real observations with theoretical ones created through simulation. To simulate real observations, a virtual Velodyne scanner is used, which is placed in the same position and orientation as the real one. Theoretical observations for the virtual scanner are created based on the known mechanism of scanner operation and an accurate and very dense terrestrial laser scanning point cloud. Experiments executed for the Velodyne HDL-32E scanner proved that the accuracy of distance measurement is comparable to that provided by the manufacturer, but different for different laser diodes, while the accuracy of horizontal angle measurement is equal to about 0.04°. Moreover, it was shown that the scanner's rotation frequency, which determines the value of the horizontal angle, is different from the nominal value and is not constant during the entire rotation. The developed observation simulation method can be used in the future to calibrate similar scanners of this type.

20

Prediction of preliminary tests for cervical cancer using artificial intelligence models

Mohanraj V., Senthilkumar J., Suresh Y., Valaramathi B., Sivanantham S.

Przegląd Elektrotechniczny

|

2024

|

R. 100, nr 7

89--95

EN

Nowadays, Aritificial Intellgience (AI) based models are extensively used in the medical science for early detection of choronic diseases. AI model plays a vital role in detecting cervical cancer in women at early stage. Cervical cancer is abnormal growth of cells in the cervix. Vagina is connected to uterus through the cervix. Mostly, various strains of Human papillomavirus (HPV) cause the infection over the cervix. A prolonged virus infection over cervix causes some cervical cells become cancer cells. It is difficult to dectect early sign of the cervical cancer. The proposed method explores cervical cancer detection and provides information on the necessary tests to be taken.The initial level of testing is achieved by getting information from users directly and processing it using a Decision Tree based classifier model. The classifier provide information on the mandatory tests that have to be taken. Then the secondary level of testing is carried out using Deep Convolution Neural Network model over a Colposcopy image of the cervix to identify the tumor region in the cervix. The model predicts the causes of cervical cancer based on the collected user information. The performance of the algorithm is evaluated based on Test accuracy, Recall, and precision. The highest cervical cancer prediction accuracy is achieved through AI model comprising Decision Tree and Deep Convolution Neural network model.

PL

Obecnie modele oparte na sztucznej inteligencji (AI) są szeroko stosowane w naukach medycznych do wczesnego wykrywania chorób kosmówkowych. Model AI odgrywa kluczową rolę w wykrywaniu raka szyjki macicy u kobiet we wczesnym stadium. Rak szyjki macicy to nieprawidłowy rozrost komórek szyjki macicy. Pochwa jest połączona z macicą poprzez szyjkę macicy. Zakażenie szyjki macicy powodują głównie różne szczepy wirusa brodawczaka ludzkiego (HPV). Długotrwała infekcja wirusowa szyjki macicy powoduje, że niektóre komórki szyjki macicy stają się komórkami nowotworowymi. Trudno jest wykryć wczesne objawy raka szyjki macicy. Proponowana metoda bada wykrywanie raka szyjki macicy i dostarcza informacji na temat niezbędnych badań, które należy wykonać. Początkowy poziom badań osiąga się poprzez bezpośrednie uzyskanie informacji od użytkowników i przetworzenie ich przy użyciu modelu klasyfikatora opartego na drzewie decyzyjnym. Klasyfikator dostarcza informacji na temat obowiązkowych badań, które należy wykonać. Następnie przeprowadza się drugi poziom badań, wykorzystując model sieci neuronowej o głębokim splocie na podstawie obrazu szyjki macicy z kolposkopii w celu zidentyfikowania obszaru nowotworu w szyjce macicy. Model przewiduje przyczyny raka szyjki macicy na podstawie zebranych informacji od użytkownika. Wydajność algorytmu ocenia się na podstawie dokładności testu, przypomnienia i precyzji. Najwyższą dokładność przewidywania raka szyjki macicy osiąga się dzięki modelowi AI obejmującemu drzewo decyzyjne i model sieci neuronowej Deep Convolution.