Viktor Nagy

Current research and development in understanding road users’ driving behaviors play a key role in improving traffic safety. Recently, several driving simulators have been employed as a suitable approach to investigate several drivers’ responses in challenging traffic scenarios. Although professional drivers represent a particular category among driving populations, the body of literature about their comparative behavioral and psychological characteristics remains limited. This study examined the differences in driving performance and visual and physiological responses between civilian and professional drivers in a simulated environment. A total of 30 drivers, with an equal split between professional and civilian categories, took part in a series of driving simulations. The simulations incorporated various infrastructure types, including four cone avoidance tasks and a high-speed motorway task. This study collected comprehensive data on performance metrics, hand usage, heart rate, and eye movements. Eye-tracking technology was used to measure visual attention. The findings revealed that during cone avoidance scenarios, civilian drivers exhibited a similar performance, visual behavior, and physiological response, except for the speed, experiment duration, and throttle, to professional drivers. In the motorway scenario, all metrics showed no significant difference between the two driver groups. These results highlight the need for cautious interpretation, particularly given the limitations of the sample. Revalidation is needed in larger studies, especially for understanding the differences between drivers’ metrics, which is crucial to elevate drivers’ safety, and assessing training programs in Hungary.

This study aims to enhance safety within the Budapest tram network by developing methods to assess and manage tram drivers’ cognitive workload. While defensive driving reduces accident risk, it becomes less effective when drivers are mentally overloaded. There is currently no reliable method to objectively measure this workload, which this research aims to develop. Trams frequently interact with unpredictable road users, increasing the likelihood of sudden braking and related injuries. Using accident data, high-risk locations were identified, and cognitive workload was assessed via eye-tracking (blinks and fixations) in a tram simulator. Participants drove two predefined routes and responded to unexpected events as they would in real traffic. Results reveal a correlation between blink/fixation frequency and cognitive load, enabling the identification of mentally demanding locations. These insights support targeted interventions to reduce driver workload and enhance operational safety.

Enhancing road safety requires an accurate assessment of the drivers’ mental states. The Driver Status Test App (DSTA) is designed to detect conditions such as intoxication, fatigue, and cognitive impairment in simulated driving environments. Utilizing a dual-platform approach, it integrates mobile data collection via React Native and Firebase with web-based management using React and TypeScript. The mobile application conducts real-time assessments of cognitive and motor functions, while the web interface offers data visualization, trend analysis, and results exportation. DSTA evaluates driver impairment through metrics such as tracking, precision, balance, and choice reaction, producing an objective impairment score. These assessments are rapid, scalable, and adaptable for various research and regulatory purposes. The composite scoring framework differentiates between impaired and unimpaired states, making DSTA valuable for driver training programs, regulatory assessments, and autonomous vehicle research, where monitoring human factors is crucial.

Driver distraction can have severe safety consequences, particularly in public transportation. This paper presents a novel approach for detecting bus driver actions, such as mobile phone usage and interactions with passengers, using Kolmogorov–Arnold networks (KANs). The adversarial FGSM attack method was applied to assess the robustness of KANs in extreme driving conditions, like adverse weather, high-traffic situations, and bad visibility conditions. In this research, a custom dataset was used in collaboration with a partner company in the field of public transportation. This allows the efficiency of Kolmogorov–Arnold network solutions to be verified using real data. The results suggest that KANs can enhance driver distraction detection under challenging conditions, with improved resilience against adversarial attacks, particularly in low-complexity networks.

The advent of autonomous vehicles (AV) could revolutionize the automotive industry by significantly improving safety, efficiency, and accessibility. Despite their potential to improve traffic safety by reducing human error, their integration into existing transportation systems presents significant challenges. This is particularly evident in scenarios involving takeover events, where there is a transition of control from the vehicle to the human driver. Our driving simulator study, involving 14 drivers in a work-zone environment, provides critical insights into the takeover performance of level 3 to level 5 AVs. The findings suggest that the successful integration of AVs depends on their seamless incorporation into existing systems and the readiness of drivers to adapt to this emerging technology.

The Automated Driving Data Analyzer (ADDA) is a modular application written in Python for the synchronized acquisition and analysis of physiological, vehicle, and interface data. It provides a managed data acquisition process and one-click data analysis. It also provides raw data storage and systematic archiving of data sets. ADDA integrates real-time data from a BeamNG vehicle simulation, Pupil Labs eye-tracking system, hand-tracking, and cardiac data. This integration allows the simultaneous recording and analysis of multiple data streams, which can be visualized and controlled through a graphical user interface (GUI) built with Tkinter. The application is designed to help researchers and engineers analyze driving behavior under different conditions, enabling a deeper understanding of the interactions between the driver and automated driving functions.

This research introduces a new approach for detecting mobile phone use by drivers, exploiting the capabilities of Kolmogorov-Arnold Networks (KAN) to improve road safety and comply with regulations prohibiting phone use while driving. To address the lack of available data for this specific task, a unique dataset was constructed consisting of images of bus drivers in two scenarios: driving without phone interaction and driving while on a phone call. This dataset provides the basis for the current research. Different KAN-based networks were developed for custom action recognition tailored to the nuanced task of identifying drivers holding phones. The system’s performance was evaluated against convolutional neural network-based solutions, and differences in accuracy and robustness were observed. The aim was to propose an appropriate solution for professional Driver Monitoring Systems (DMS) in research and development and to investigate the efficiency of KAN solutions for this specific sub-task. The implications of this work extend beyond enforcement, providing a foundational technology for automating monitoring and improving safety protocols in the commercial and public transport sectors. In conclusion, this study demonstrates the efficacy of KAN network layers in neural network designs for driver monitoring applications.

Summary. In the area of road safety, the development of a Modular, Machine Vision-Based, Custom-Built Driver Monitoring System (DMS) for bus drivers has become imperative. This research presents a comprehensive system capable of detecting drowsiness, blinking patterns, and various forms of distraction, including the use of mobile phones, and one-handed driving. Leveraging the power of Mediapipe and YOLOv7 for real-time image analysis, as well as ROS2 for seamless data transfer, our system not only ensures the immediate safety of bus passengers but also offers expandable functionality, such as eye tracking and skeleton detection. Összefoglalás. A biztonságos tömegközlekedés iránti igény fokozza a járművezetőkre nehezedő nyomást a növekvő forgalmi torlódások miatt. Ezt súlyosbítja a közlekedési rendszer összetettsége és a fokozódó külső ingerek hatása, különösen városi környezetben. A közösségi közlekedésben alkalmazható, a közlekedésbiztonságot fokozó moduláris, gépi látáson alapuló, egyedi fejlesztésű járművezető-felügyeleti rendszer kifejlesztése az autóbuszvezetők monitorozására elengedhetetlen. Ez a kutatás egy olyan átfogó rendszert mutat be, amely képes érzékelni az álmosságot, a pislogási mintákat és a figyelemelterelés (disztrakció) különböző formáit, beleértve a mobiltelefon-használatot, és az egykezes vezetést. A Mediapipe és a YOLOv7 valós idejű képelemzésre, valamint a ROS2 adatátvitelre való felhasználásával rendszerünk nemcsak a busz utasainak biztonságát garantálja, hanem olyan bővíthető funkcionalitást is kínál, mint például szemkövetés és csontvázfelismerés (szkeleton). A rendszer alapvető célja az, hogy a szemmozgás, fejtartás és testtartás elemzésével pontosan azonosítja a járművezető fáradtságát, figyeli a pislogási mintákat az álmosság jeleit, és felismeri a közúti biztonságot veszélyeztető disztrakciókat. Továbbá a rendszer moduláris felépítése lehetővé teszi további érzékelők, például szemmozgás-követő rendszer, telemetriai eszközök vagy 5G-adapterek egyszerű integrálását, ami átfogó megfigyelést és adatfúziót tesz lehetővé a valós környezetbe történő adaptálás elősegítésére. A fejlesztett Járművezető Monitoring Rendszer a ROS2 keretrendszer segítségével integrált megoldást kínál a buszvezetők megfigyelésére. A rendszer alapvető képessége a fedélzeti kamerák által rögzített felvételeken a járművezető vizuális felismerése. Ezen túlmenően képes követni a vezető testén lévő kulcspontok, mint a fej, a törzs és a karok pozícióját, ami létfontosságú az ő testtartásának és mozgásának megértésében. Az egyediséget a vezetőfülke mérete és a buszvezető személygépjárműhöz mérten dinamikus mozgásképe adja. A rendszer az emberi arcot is részletesen elemzi, kiemelve a fontos arcpontokat, mint a szemek, orr és száj. Ez lehetővé teszi a tekintet irányának, arckifejezéseknek, valamint a fáradtság vagy stressz jeleinek azonosítását. Az adatok könnyebb értelmezése érdekében a rendszer egy vizuális ábrázolást is nyújt az észlelési folyamatról. Az adatok kezelése és kommunikációja a ROS2 keretrendszeren keresztül történik, amely strukturált módon rendezi az adatokat és támogatja a valós idejű feldolgozást és elemzést. Az összegyűjtött adatok tárolására a .rosbag fájlformátumot használjuk, amely lehetővé teszi az adatok hatékony rögzítését és későbbi felhasználását elemzésekhez és felülvizsgálatokhoz. A tanulmány a moduláris járművezető-felügyeleti rendszer felépítését, megvalósítását és tesztelését mutatja be, részletesen közli az alkalmazott algoritmusokat és technológiákat. A valós körülmények között végzett kísérletek eredményei bizonyítják a rendszer hatékonyságát, valamint a rendszer szélesebb közlekedési ökoszisztémákba való integrálhatóságát. A buszvezető monitorozása kapcsán kapott adatok hozzájárulhatnak a jármű és utasai biztonságának fokozásához. A járművezető figyelmének nyomon követése és a kognitív terhelés elemzése lehetőséget kínál a munkakörülmények optimalizálására és a balesetmegelőzési megoldások javítására.

On-road driving studies are essential for comprehending real-world driver behavior. This study investigates the use of eye-tracking (ET) technology in research on driver behavior and attention during Controlled Driving Studies (CDS). One significant challenge in these studies is accurately detecting when drivers divert their attention from crucial driving tasks. To tackle this issue, we present an improved method for analyzing raw gaze data, using a new algorithm for identifying ID tags called Binarized Area of Interest Tracking (BAIT). This technique improves the detection of incidents where the driver’s eyes are off the road through binarizing frames under different conditions and iteratively recognizing markers. It represents a significant improvement over traditional methods. The study shows that BAIT performs better than other software in identifying a driver’s focus on the windscreen and dashboard with higher accuracy. This study highlights the potential of our method to enhance the analysis of driver attention in real-world conditions, paving the way for future developments for application in naturalistic driving studies.