TY - JOUR AU - Tollner, Dávid AU - Ziyu, Wang AU - Zöldy, Máté AU - Török, Árpád TI - Demonstrating a new evaluation method on ReLU based Neural Networks for classification problems JF - EXPERT SYSTEMS WITH APPLICATIONS J2 - EXPERT SYST APPL VL - 250 PY - 2024 PG - 12 SN - 0957-4174 DO - 10.1016/j.eswa.2024.123905 UR - https://m2.mtmt.hu/api/publication/34774286 ID - 34774286 AB - Deep neural networks, which have proven to be effective methods to solve complex problems, can even be applied in decision systems controlling critical processes. However, in such applications the outcomes of the neural network must be checked if it we have a clear understanding regarding the operation of network at given input intervals. The most straightforward, though often computationally expensive approach for this checking process evaluates the network at discrete input points and estimates the expected outputs at the given interval. The present research aims to develop a novel approach that can identify in the case of specific input intervals whether the operation process and the output of a neural network can be considered as known or unknown. During the presented case study, we investigated the ReLU (Rectified Linear Unit) and Sigmoid activation functions, using the double moon and the Banknote Authentication classification problems for demonstration. Our method can be applied to identify certain input intervals where the given neural network cannot support critical decisions. The evaluation is performed based on a nonlinear system of equations and inequalities built on arbitrary continuous activation functions. To define the critical intervals of the input variables (i.e. where the decision-making system should not be relied on), those input variable combinations are identified which result in a non-expected output value. This inverse logic is intended to identify intervals of the input variables where the response of the system and the correct decision are not identical. The presented demonstration examples supported our assumption that the number of the neurons and the dimension of the decision space have a significant impact on the complexity of the evaluation process. © 2024 The Author(s) LA - English DB - MTMT ER - TY - JOUR AU - Elaa, Elgharbi AU - Zöldy, Máté AU - Safa, Bhar Layeb TI - Autonomous Vehicle and Pedestrian Interaction. Leveraging The Use of Model Predictive Control & Genetic Algorithm TS - Leveraging The Use of Model Predictive Control & Genetic Algorithm JF - COGNITIVE SUSTAINABILITY J2 - COGSUST VL - 3 PY - 2024 IS - 1 SP - 15 EP - 31 PG - 17 SN - 2939-5240 DO - 10.55343/cogsust.90 UR - https://m2.mtmt.hu/api/publication/34774194 ID - 34774194 LA - English DB - MTMT ER - TY - JOUR AU - Ferencz, Csanád AU - Zöldy, Máté TI - Neural Network-based Multi-Class Traffic-Sign Classification with the German Traffic Sign Recognition Benchmark JF - ACTA POLYTECHNICA HUNGARICA J2 - ACTA POLYTECH HUNG VL - 21 PY - 2024 IS - 7 SP - 203 EP - 220 PG - 18 SN - 1785-8860 DO - 10.12700/APH.21.7.2024.7.12 UR - https://m2.mtmt.hu/api/publication/34746683 ID - 34746683 LA - English DB - MTMT ER - TY - JOUR AU - Zöldy, Máté AU - Baranyi, Péter Zoltán AU - Török, Ádám TI - Trends in Cognitive Mobility in 2022 JF - ACTA POLYTECHNICA HUNGARICA J2 - ACTA POLYTECH HUNG VL - 21 PY - 2024 IS - 7 SP - 189 EP - 202 PG - 14 SN - 1785-8860 DO - 10.12700/APH.21.7.2024.7.11 UR - https://m2.mtmt.hu/api/publication/34746682 ID - 34746682 LA - English DB - MTMT ER - TY - JOUR AU - Virt, Márton AU - Zöldy, Máté TI - Enhancing the Viability of a Promising E-Fuel: Oxymethylene Ether–Decanol Mixtures JF - ENERGIES J2 - ENERGIES VL - 17 PY - 2024 IS - 6 PG - 17 SN - 1996-1073 DO - 10.3390/en17061348 UR - https://m2.mtmt.hu/api/publication/34734773 ID - 34734773 N1 - Export Date: 5 April 2024 Correspondence Address: Zöldy, M.; Department of Automotive Technologies, Műegyetem rkp. 3., Hungary; email: zoldy.mate@kjk.bme.hu AB - Achieving sustainable mobility is a crucial factor in maintaining long-term economic growth without adverse effects on human health and the environment. E-fuels, such as the promising oxymethylene ether (OME), can contribute to sustainable road transport. However, this compound does not meet the requirements of EN590; thus, it is unsuitable for unmodified diesel engines. This work aims to improve the applicability of OME by blending it with n-decanol, which can also be produced sustainably. Combustion and emissions were investigated in a medium-duty commercial diesel engine with different binary and ternary mixtures of oxymethylene ether, n-decanol, and B7 diesel. Laboratory analysis of six key mixture parameters revealed that the formulated blends met the EN590 requirements, with the exception of that of density. The results demonstrated that the created mixtures, including one without any diesel fuel, can be efficiently utilized in unmodified diesel engines. OME’s beneficial effects on combustion and emission were preserved while its viability was improved; a maximum increase of 7.6% in brake thermal efficiency was observed, alongside a potential decrease of nearly 70% in PM emissions at unaltered NOx levels. LA - English DB - MTMT ER - TY - JOUR AU - Virt, Márton AU - Zöldy, Máté TI - Cost Efficient Training Method for Artificial Neural Networks based on Engine Measurements JF - ACTA POLYTECHNICA HUNGARICA J2 - ACTA POLYTECH HUNG VL - 21 PY - 2024 IS - 7 SP - 123 EP - 145 PG - 23 SN - 1785-8860 DO - 10.12700/APH.21.7.2024.7.8 UR - https://m2.mtmt.hu/api/publication/34730296 ID - 34730296 LA - English DB - MTMT ER - TY - JOUR AU - Ferencz, Csanád AU - Zöldy, Máté TI - Exhaustive Investigation of the Promises and Perils of Autonomous Mobility Technology JF - PERIODICA POLYTECHNICA TRANSPORTATION ENGINEERING J2 - PERIOD POLYTECH TRANSP ENG VL - 52 PY - 2024 IS - 1 SP - 86 EP - 95 PG - 10 SN - 0303-7800 DO - 10.3311/PPTR.22573 UR - https://m2.mtmt.hu/api/publication/34498268 ID - 34498268 N1 - Export Date: 12 January 2024 CODEN: PPTED Correspondence Address: Ferencz, C.; Department of Automotive Technologies, Stoczek J. u. 6., Hungary; email: csanadferencz@edu.bme.hu Funding details: Budapesti Műszaki és Gazdaságtudományi Egyetem, BME Funding text 1: The realization of this present scientific research project would not have been possible without the support of the Department of Automotive Technologies of the Budapest University of Technology and Economics. AB - The technology of autonomous driving systems has the potential to bring significant benefits in terms of safety, mobility, and congestion, as well as land use and energy consumption. In this paper, the authors present both the promises and possible dangers of the large-scale implementation of autonomous vehicle technologies in the current road traffic networks. The effects of autonomous mobility technology are complex and require careful consideration. While there is great potential to transform transportation and offer significant improvements in many areas, there are also risks and challenges that need to be addressed to ensure safe, efficient, and equitable deployment of the technology. © 2024 Budapest University of Technology and Economics. All rights reserved. LA - English DB - MTMT ER - TY - CHAP AU - Tollner, Dávid AU - Zöldy, Máté TI - Dynamic Data extraction from vehicles through OBD connector T2 - 2023 IEEE 2nd International Conference on Cognitive Mobility (CogMob) PB - IEEE CY - Piscataway (NJ) PY - 2023 SP - 189 EP - 194 PG - 6 UR - https://m2.mtmt.hu/api/publication/34414615 ID - 34414615 LA - English DB - MTMT ER - TY - CHAP AU - Ferencz, Csanád AU - Zöldy, Máté TI - Failure rate analysis of driverless car reliability using different confidence intervals T2 - 2023 IEEE 2nd International Conference on Cognitive Mobility (CogMob) PB - IEEE CY - Piscataway (NJ) PY - 2023 UR - https://m2.mtmt.hu/api/publication/34402144 ID - 34402144 LA - English DB - MTMT ER - TY - GEN AU - Nyerges, László Ádám AU - Zöldy, Máté TI - Speed profile generation for an electric race vehicle PY - 2023 PG - 2 UR - https://m2.mtmt.hu/api/publication/34220816 ID - 34220816 LA - English DB - MTMT ER -