Műegyetemi Digitális Archívum

Conference on BME ZalaZONE 2022

Permanent URI for this collection


Recent Submissions

Now showing 1 - 22 of 22
  • Item
    Testing and Validation of the FRC08 Formula Student Race Car at the ZalaZONE Automotive Test Track
    (2022) Farkas, Zsolt
    The development and adjustment process of a race car consists of countless steps that require a lot of decisions to be made. These decisions can only be effective if we have the right amount and quality of information. A significant part of this information is provided by computer simulations, which, unfortunately, often do not cover reality due to the inaccuracy and error of the model setup. A simulation can be considered credible if certain elements of it can be replicated in reality or give almost the same results. In order to map the characteristics of the tire on the FRC08 car of the BME Motorsport Formula Student racing team, we carried out tire temperature measurements. The tests were carried out at the ZalaZONE automotive test track, which provides a unique opportunity and working environment internationally for testing and validation of simulation models. Getting to know the behaviour of the race car in as much detail as possible also revealed the strengths and weaknesses of the vehicle. This provides the team with the opportunity to adjust the optimal operating parameters of the vehicle and provides appropriate information for the development of a new racing car.
  • Item
    Collaboration Possibilities for Autonomous Industrial Transport Vehicles
    (2022) Bohács, Gábor; Horváth, András Máté
    Autonomous industrial transport vehicles are already in operation in the industry, but the area is very rich in further development topics. Collaboration among these vehicles makes up a special problem. First, this paper surveys the current area in detail. There is not only collaboration among the same type of vehicles like platooning discussed, but also the special tasks that are necessary for material handling operations. Second, proper conclusions are drawn for the elaborated concept of our former research regarding the possibilities of enhancement towards vehicle collaboration and its conditions.
  • Item
    Control Design of an Autonomous Moving Platform for Test Tracks
    (2022) Kocsány, László; Markovits, Gergely Tibor
    In recent years, artificial intelligence, deep learning, and computer vision systems have paved the way for the development of various self-driving vehicles. The risk of testing these vehicles is not negligible, given the high kinetic energy, so testing methods should be chosen carefully. In addition to static objects, the problem of handling dynamic objects during the test cases which are carried out on test tracks as ZalaZONE. Hence dynamic objects are carried by self-driving platforms that do not cause significant material damage to either the test device or the test subject in the event of a loss of control. This paper presents two important aspects of the development of a universal Moving Platform. These are the safety analysis of the onboard architecture to ensure a highly reliable emergency stop and the trajectory tracking methods and the implementation of the associated low-level control loops.
  • Item
    Predictive Maintenance in Distributed Environments
    (2022) Sik, Dávid; Ekler, Péter; Levendovszky, János
    In this paper we present a concept to implement predictive maintenance in distributed environments. A distributed environment can involve several problems to be solved such as the different speed and frequency of data, the different priorities of data sources, the different dimensions and data structures and also the volume of the data to be processed. There are software, frameworks, algorithms to work out these problems, however the interoperability is harder to maintain between these. Recently it become possible to use historical and real-time datasets of parameters and key performance indicators of environments in order to prevent failures and monitor the system state. The aim is to outline a solution, using open-source tools, to support the lifecycle from the data extraction, through the transformation and algorithmic steps, until the usage and visualization of the gained information and feedback of the data to the environment.
  • Item
    Conflict Management Algorithms Development Using the Automated Framework for Autonomous Vehicles
    (2022) Sziroczák, Dávid; Rohács, Dániel
    The fields of autonomous ground and aerial vehicles are a young, dynamically expanding field of industry. The current traffic management solutions are inadequate to support the predicted high volume of operations efficiently and safely. As such, novel, automated management solutions are required. This paper presents the research work performed on the automated conflict management prototype system, combining conflict management for autonomous ground and aerial vehicles. This paper briefly presents the background and architecture of the system prototype developed, then highlights the research directions related to the conflict management algorithm development. Algorithms need to provide the desired amount of safety, while at the same time ensuring acceptable levels of performance, cost, and efficiency. Potential conflict management algorithms are investigated, and a computational methodology developed for objective comparison, to enable the future, data-based development of the conflict management framework.
  • Item
    Establishment of a Local GNSS Correction Service for the Localization of Autonomous Vehicles
    (2022) Rózsa, Szabolcs; Ács, Ágnes; Turák, Bence
    Accurate localization of autonomous vehicles is a key component of the onboard control and guidance system. Global Navigation Satellite Systems (GNSS) are widely used in the transportation industry for positioning and navigation
  • Item
    Routing Algorithms for Wireless Sensor Networks in Smart Cities
    (2022) Pásztor, Dániel; Ekler, Péter; Levendovszky, János
    Wireless Sensor Networks (WSNs) are one of the most important parts of the advancements in smart city planning. The sensors provide valuable data about various metrics
  • Item
    Double Lane Change Path Planning Using Reinforcement Learning with Field Tests
    (2022) Fehér, Árpád; Aradi, Szilárd; Bécsi, Tamás
    Performing dynamic double lane-change maneuvers can be a challenge for highly automated vehicles. The algorithm must meet safety requirements while keeping the vehicle stable and controllable. The problem of path planning is numerically complex and must be run at a high refresh rate. The article presents a new approach to avoiding obstacles for autonomous vehicles. To solve this problem
  • Item
    Point Cloud Based Road Surface Modelling and Assessment
    (2022) Lovas, Tamás; Baranyai, Dániel; Somogyi, Árpád
    Creating digital twins of the built environment based on point clouds broadens its application area; point clouds of buildings support BIM (Building Information Modeling) while the digital twins of road surfaces support transportation applications. Point clouds acquired by current TLS (Terrestrial Laser Scanning) or by MLS (Mobile Laser Scanning) systems represent the road surface with high accuracy, and resolution (i.e. with small point spacing). In particular applications extreme high accuracy and robustness is required; the paper discusses surveying and assessment of vehicle test track, including a braking platform. Both the surveying method (TLS supported by Total Station measurements) and the data processing workflow are presented. The potential evaluation results are discussed in detail; semi-automatically created 2D sections and deviation maps shows how the current state of the pavement differs from the as-designed geometry. The investigations proved that point-cloud based data acquisition methods enable deriving high accuracy and high-resolution road surface models.
  • Item
    Automotive Proving Ground HD Map Models
    (2022) Panker, Dániel; Tettamanti, Tamás
    Self-driving vehicles also need maps - just like us drivers - for route plans, but traditional maps are often inaccurate and contain significantly less information than so-called High Definition Maps (HD). HD maps are usually based on laser measurements, thus the accuracy of each road element can be around 2 centimeters. Our goal is to properly support research in this direction as well, therefore we created an HD map of the selected modules of the ZalaZONE proving ground. The selected components are the High-Speed Handling course, Dynamic Platform, ADAS surface and Motorway section. These materials are accessible for download including the high-resolution and high-precision Lidar point cloud which the maps are based on, recorded with the Lecia Pegasus 2 system. We have implemented our HD maps in the increasingly widespread OpenDRIVE format in the automotive industry.
  • Item
    Design of a Novel Road Traffic Control System for ZalaZONE Proving Ground
    (2022) Tettamanti, Tamás; Varga, István
    The testing of Connected and Automated Vehicles (CAVs) and that of the smart infrastructure (traffic control devices and vehicle sensors) in relation with CAVs will be supported by the development of a novel type of road traffic light management system at ZalaZONE Automotive Proving Ground. The system to be developed aims to allow a fully flexible traffic signal control during vehicle and system testing. The system shall provide a freely programmable open Application Programming Interface (API) towards the traffic light control units in contrast with traditional (rigid and closed) traffic control equipments. In this concept
  • Item
    Test Automation – From Virtual Scenario Creation to Real-world Testing
    (2022) Horváth, Áron; Tettamanti, Tamás
    As advanced driver assistance systems (ADAS) are continuously upgrading, the testing standards, descriptions are evolving as well. The border disappears more and more between simulation and real-world testing. For more accurate repeatable scenarios we need a new method, which helps to make faster the process of scenario creation. The result for this would be the connection of simulated scenarios and real test equipment. Our result includes several components such as HD maps, industrial simulation software, OpenSCENARIO compatibility, MATLAB programming, AB Dynamics software and test equipment. With HD maps and OpenSCENARIO compatible simulation software we can create GPS accurate scenarios. Then, the own developed MATLAB program can convert the OpenSCENARIO file into a new format, which is readable by AB Dynamics software. After this process the last step is the real-world testing with AB Dynamics test equipment.
  • Item
    Initiation and Stabilization of Drifting Motion of a Self-driving Vehicle with a Reinforcement Learning Agent
    (2022) Tóth, Szilárd Hunor; Bárdos, Ádám; Viharos, Zsolt János
    Performing special driving techniques like drifting can be challenging even for professional human drivers. However, such maneuvers can be essential for avoiding accidents in critical road scenarios like evasive maneuvers. This paper reports novel research results whose main goal is to develop a self-driving agent for drift motion control based on vehicle simulation in MATLAB/Simulink. The state representation of the vehicle includes the longitudinal and lateral velocities with the yaw rate. The agent action space consists of two actuators: the throttle position and the roadwheel angle. The goal of the agent is twofold: first, it needs to jump into a drifting state; second, it has to keep the vehicle in drift. The simulation results show that the proposed RL agent is capable of learning to approach a predetermined drift equilibrium from cornering and staying in this drift situation as well. For the training, the solution excluded using any prior data. It only works with information gained from the simulation model, which is a remarkable difference from the actual state-of-the-art RL-based solutions.
  • Item
    Development of a GNSS Based High Accuracy Measurement System to Support Vehicle Dynamics Testing
    (2022) Fehér, Árpád; Kardos, Gábor; Szabó, Ádám; Aradi, Szilárd
    The article presents the development of a low-budget positioning device that aims to provide an alternative in self-driving vehicle development research that could replace costly, commercially available devices. In addition to being financially advantageous, it has the added benefit of allowing students to be involved in development. The primary function of the device is the sensor fusion, which outputs position, velocity, and orientation estimation based on data provided by Real-Time Kinematic (RTK) Global Navigation Satellite System (GNSS) technology and an Inertial Measurement Unit (IMU). High-frequency estimates are generated by running an Extended Kalman Filter (EKF) on a microcontroller in an embedded environment. During the work, new challenges arose several times that required solutions. For example, delays due to the operation of GNSS receivers, which the estimation algorithm must compensate, and proper calibration of the sensors for the measurement vehicle. In addition to the software, the development of the tool includes the complete design, manufacture, and testing of the hardware, which allows testing the completed software units not only in a simulation but also in a real environment. During testing, the output of the developed device was compared several times with commercially available hardware for similar purposes.
  • Item
    3D Object Detection in LIDAR Point Cloud Based on Background Subtraction
    (2022) Nagy, Szabolcs; Rövid, András
    Autonomous vehicles have a key role in transportation systems of the future, but there are still many difficulties to overcome. Nowadays one of the most critical problems in autonomous driving is the precise and robust detection of traffic participants. This paper presents a LIDAR-based 3D object detection method. The algorithm uses HD Map to subtract the static background points from the LIDAR point cloud. The remaining points are grouped by clustering, then 3D boxes are fitted to the clusters. The object detection method presented in this paper was tested on real sensor data collected by a solid-state LIDAR on the highway module of the ZalaZONE proving ground. The results showed that the developed algorithm performs as intended in a highway scenario, detecting vehicles even more than 100 meters away from the sensor by a framerate of ~20FPS.
  • Item
    Simulation of Road Traffic Accidents Related to ADAS Systems in PreScan
    (2022) Kazár, Tamás Márton; Pethő, Zsombor; Vida, Gábor; Török, Árpád
    This paper describes a simulation of a specific accident reconstructed in PreScan software. Through this, we show that the PreScan software can be efficiently applied in accident reconstruction, particularly the operation of ADAS systems can be explored and analyzed. Beyond this, we also demonstrated during our research that the investigated simulation environment can be used efficiently to evaluate the reliability and safety of electronic control units developed to control highly automated vehicle functions. Consequently, we also proved the applicability of concept related to the hybrid multi-agent simulation environment (especially considering the ZalaZONE ecosystem) including differently controlled components (e.g. fully autonomous, human-driven or highly automated vehicles). In the simulation, we examine the Adaptive Cruise Control (ACC), Lane Keeping Assist (LKA), and Automatic Emergency Braking System (AEBS) in the accident. Evaluating the results, we found that the lane-keeping system is very sensitive to the quality of pavement signals. In the selected case, poor-quality signs likely played a major role. On the other hand, it was clear that the properly working safety systems would have been effective in reducing collisions speed, and a longer tracking distance would have had a positive effect. In the case of the adequate operating FCW, the driver would have had sufficient time to intervene and stop the vehicle. In the final section, we analyzed how PreScan can be used to connect an external device to the simulation environment.
  • Item
    Teleoperation with a Real-time Digital Twin
    (2022) Remeli, Viktor; Boronyák, Ádám; Tihanyi, Viktor
    One of the state-of-the-art trends in intelligent transportation systems research is the development of real-time digital twin technologies and their applications for increased safety and autonomy on the roads. The areas of deployment also include test tracks, logistic yards, factories, hospitals, airports, railways, and any such constrained and managed environment where transportation automation may be of benefit. While applying the same principles as for open road networks, managed environments allow the efficient deployment of optimally distributed sensory infrastructure while also observing certain simplifying assumptions (constraints on the expected scenarios), both factors leading to earlier adoption and ROI. In the case of fleets or swarms of automated vehicles, certain cost savings can be enacted by recovering the environmental model exclusively from infrastructure sensors. However, such a setup is never exempt from the necessity of occasional manual intervention performed by a teleoperator. In our current paper we explore system and HMI designs that would facilitate safe and efficient on-demand teleoperation functionality in such a system.
  • Item
    On-Site Test Measurements at ZalaZONE Automotive Proving Ground for Aiding Various Research Projects
    (2022) Vincze, Zsolt; Rövid, András
    The ZalaZONE Automotive Proving Ground can provide multiple controlled and safe test environments which becomes essential when various measurements and field tests are required for different vehicle-related scientific research projects. The Motorway and Smart City platforms are detailed models of a real highway section and a densely populated city area. These test sites were used to set up multiple field tests for various Intelligent Transportation Systems related research projects we are working on. Various measurements were performed by deploying multiple sensors of different types along the Motorway module as well as in the SmartCity zone in form of sensor stations. The Smart City platform was used to perform pre-accident specific scenarios and collect corresponding data in order to support the development of methods which are aiming for the recognition of traffic situations where the occurrence of an accident is highly probable.
  • Item
    Model Predictive Controller for Path Following Applications
    (2022) Domina, Ádám; Tihanyi, Viktor
    In this article, a model predictive controller (MPC) for automated vehicle path following is presented. The MPC calculates the optimal steering command based on the prediction of the future states of the vehicle and the reference given for the controller. The aim is to minimize the difference between the predicted and the reference states, which is ensured by the optimal control input. The MPC control technique applies a vehicle model for state prediction. In this paper, a bicycle model is used for state prediction, which model is updated at every time step formalizing an LPV-MPC structure. The controller aims to minimize the lateral and angular deviation of the vehicle from a fixed path. A reference path is built, on which the controller is tested, showing a good performance.
  • Item
    Survey on Image Based Object Detectors
    (2022) Cserni, Márton; Rövid, András
    Sensor fusion-based detector utilizes camera sensors to solve the problem of recognizing objects and their classifications accurately. This has been proven to increase accuracy compared to single sensor detectors and can significantly help with the 3D tracking of vehicles in the sensor system’s area of interest. Even at a distance, where no lidar points are available from the target, a high-resolution camera-based detector can easily detect and classify vehicles. There is a variety of real-time capable 2D object detector convolutional neural networks, some of which are open source. This survey compiles a list of these algorithms, comparing them by precision scores on well-known datasets, and based on experimental evaluation completed on camera images taken on the ZalaZONE test-track to evaluate the distances at which the detectors first perceived the test vehicles. Additionally, inference times are also compared.
  • Item
    Development of 3D Visualization for Displaying Real-time Sensor Information from Multiple Sources
    (2022) Szalai, Mátyás; Remeli, Viktor; Tihanyi, Viktor
    For modern scientific results, it is increasingly challenging to present them in a form that the general public can understand. This is particularly true in the field of collaborative intelligent transportation services and connected vehicles, where many different platforms and sensors are operating simultaneously. It is important for users to build up the necessary trust in the new technology, which is why the presentation of the emergence and interaction of collective, platform- and sensor-level environment models is an important area. In this paper, a visualization environment that can present the different locally detected and centrally fused objects in real time are presented through an example of multiple infrastructure sensors. By statically and dynamically representing the digital twin, we can create a world where human users can see their environment through the eyes of individual vehicles, infrastructure stations, or the whole system. This, combined with Mixed Reality technology, can produce a result that feels real to the human eye and facilitates intuitive understanding of and trust towards the system.
  • Item
    A Raw Fusion Based 3D Object Detector for Pedestrian and Vehicle Position Estimation
    (2022) Csonthó, Mihály; Rövid, András
    Robust sensing of the environment is an essential and safety-critical part of self-driving vehicles. Most of the algorithms used for this problem employ sensor fusion to increase the reliability of sensing. The most common is the object-level fusion, where separate sensor detections are combined to create an object list. Less common are low-level fusion algorithms that compile their detection list from fused low-level input data. The algorithm presented here uses low-level (raw) fusion and can remarkably improve the detection accuracy and reliability compared to single-camera systems. The proposed detector is aimed for pedestrian detection but is also capable of detecting the position of vehicles or other specific objects even in cases when the number of lidar points representing the object is low.