Imitation learning has been one of the mainstream methods for achieving autonomous driving in recent years. Despite its simple structure and expert-like performance in dealing with diverse driving situations, it may lead to unsafe driving solutions due to the absence of a constraint-checking process. To resolve this problem, this paper presents a method for path planning of autonomous vehicles based on imitation learning and optimization techniques. In this paper, imitation learning is used to derive the expert-like driving path. For this, a neural network is trained in a supervised manner to derive the driving path output given environmental information and a reference path. An optimization technique is applied to post-process the imitation-based driving path result using a cost function and constraints. This contributes to enhanced safety compared to a path planner based solely on imitation learning. The proposed algorithm has been validated via simulation studies. The test results show that the proposed combination of imitation learning and optimization techniques enables the autonomous vehicle to derive a feasible driving path following the expert driving pattern. The addition of the optimization technique is shown to reduce collision rate compared to the neural network-based path planning.

In accordance with trends in responding to the global climate crisis, goals are being established to achieve carbon neutrality in various fields. One of these is a plan to convert diesel ground handling equipment at Korea's airports to eco-friendly. Currently, domestic airport ground handling companies are making efforts to convert existing diesel equipment to eco-friendly equipment, but are complaining of difficulties due to the lack of required technology and management plans. Accordingly, through research on ways to convert old diesel ground handling equipment at the airport into eco-friendly equipment, we would like to present data necessary to induce rapid eco-friendly conversion and realize the goal of carbon neutrality.

This study is a scenario in which actual tire departure accidents are reproduced using the PC-Crash program and reflected in platooning autonomous vehicles. This is a scenario in which a platooning autonomous vehicle drives in the opposite lane as a scenario in which a tire leaves and goes to the opposite lane, causing the preceding vehicle to avoid the tire and collide with the following vehicle. As a result of applying the scenario, it was possible to check whether or not it was avoided according to the inter-vehicle distance from the preceding vehicle of platooning autonomous driving. In addition, it is expected that this scenario can be used as a platooning test scenario as it can check whether or not to avoid collision through V2X communication.

This study proposes a Human-Machine Interface (HMI) implementation for autonomous vehicles aimed at ensuring the safety of vulnerable vehicle users, such as the elderly, disabled, and residents of transportation-disadvantaged areas. Particularly, sharp turn detection algorithm was developed to provide pre-emptive warnings to passengers about sudden vehicle movements, enhancing safety and comfort. Field tests conducted using high-definition map (HD map) of Sejong City demonstrated the effectiveness of the system, allowing autonomous vehicles to detect sharp turns in advance and display appropriate warning messages through the HMI. The results show that the proposed system improves the safety of vulnerable vehicle users by providing early detection of sharp turns, thus reducing potential risks during autonomous vehicle rides. This research highlights the importance of user-friendly and intuitive HMI systems for increasing public trust and acceptance of autonomous driving technologies, and it contributes to the development of a more inclusive and accessible transportation system. The findings are expected to play a key role in advancing the commercialization and widespread adoption of autonomous vehicles for all users.

According to the government's plan to complete the foundation for the commercialization of convergence level 4+ autonomous vehicles in 2027, the V2X system is expected to become basic equipment in vehicles. The V2X system broadcasts various status information, including location data, to the surrounding environment in real time. Therefore, the legal basis and installation criteria need to be established first. Based on standards, laws, administrative regulations and previous research results related to V2X system, a draft of V2X system installation standards has been derived. The draft of the Vehicle Management Act, the draft of the Enforcement Rules of the Vehicle Management Act and the draft of the Regulation on Performance and Standards of Vehicles and Vehicle Parts in relation to the installation criteria of V2X system have been studied, revised and supplemented. We believe that the draft of the installation criteria for V2X systems for the fusion of autonomous driving and C-ITS will provide a legal basis.

The essential process of automobile crash testing, which evaluates the collision safety and occupant protection performance of vehicles, is costly. As a result, virtual crash testing, which can substitute or complement real-world testing, is being researched and applied as an effective alternative. As the application scope of virtual crash testing expands, research is actively being conducted to verify the reliability of the results obtained from virtual crash tests and their similarity to real-world crash tests. In this study, the effect of changing the coefficient of friction between the passenger and vehicle contact parts during high-speed collisions in a specific range of friction coefficients was quantitatively analyzed. It was found that the impact of the friction coefficient varied depending on the analysis model, and in the crash analysis, there were differences in the maximum reaction force and the shape of the reaction force curve over time. Therefore, it is necessary to examine the range of friction coefficients applicable to virtual crash testing.

The global response to the climate crisis and the ongoing energy transition are driving a rapid restructuring of industrial systems worldwide. In the transportation sector, Battery Electric Vehicles (BEVs) and Hydrogen Fuel Cell Vehicles (FCEVs) are emerging as central technologies for achieving carbon neutrality and sustainable mobility. This study conducts a comparative analysis of BEVs and FCEVs from the perspectives of technology management and economic innovation. It further examines domestic and international policy responses and industrial strategies to assess the competitiveness and sustainability of each technology. Based on the findings, this paper offers strategic insights for securing competitive advantage in the global eco-friendly vehicle market.

This paper analyzes the relation between energy efficiency and motion sickness in electric vehicles (EVs) under different regenerative braking modes. Simulations were conducted on the FTP-75 cycle with varying braking levels. Energy consumption was evaluated, and motion sickness was measured using the ISO 2631-1 based Motion Sickness Dose Value (MSDV). Results revealed a trade-off, as higher braking levels improved energy recovery but also increased motion sickness. This quantitative analysis provides meaningful insights into the trade-off between energy efficiency and motion sickness, offering a valuable basis for future EV design and evaluation.

This study investigates the compatibility of the KADAS (Korea Automated Driving Vehicle Assessment System) - Based Vehicle in the Loop Simulation (VILS) environment with diverse ADAS Camera sensor architectures. Specifically, it analyzes the functional performance of Advanced Driver Assistance Systems (ADAS) for Vehicle A, equipped with a mono-camera system, and Vehicle B, utilizing a Stereo/Dual-camera system. The findings reveal that Vehicle A demonstrated high success rates in Adaptive Cruise Control (ACC) and Autonomous Emergency Braking (AEB) functions. In contrast, Vehicle B exhibited Severe performance degradation in the same VILS environment. This degradation is attributed to the VILS environment's inherent limitations in supporting the depth perception mechanisms of stereo/dual cameras and the conservative nature of their sensor fusion logic. This research underscores the critical need for improving ADAS evaluation technologies within VILS environments to accommodate various sensor architectures, thereby contributing essential foundational data for standardized ADAS performance assessment.

V2X communication technology can overcome the limitations of existing Lv. 3 autonomous vehicles by recognizing surrounding traffic conditions and road hazards beyond the recognition range of ADAS sensors, and its necessity is gradually increasing as a key technology for implementing Lv. 4 or higher autonomous driving. Major countries around the world (Europe, China, Japan), including Korea, are paying attention to accident prevention safety using V2X technology and have established a roadmap for the New Car Assessment Program (NCAP) and are introducing V2X technology as an NCAP evaluation item. In this study, we conducted research on the development of vehicle V2X communication performance safety evaluation technology, including evaluation items, evaluation environment, and evaluation method for evaluating the performance of congestion control technology of V2X communication devices installed in vehicles in an environment where V2X technology is applied to actual roads and numerous vehicles and infrastructures communicate simultaneously, resulting in communication congestion and V2X communication performance degradation.

This study investigates the structural characteristics and underlying risk factors of electric vehicle (EV) fires through a statistical analysis of 148 documented cases in South Korea between 2018 and 2024. While public concern regarding EV fire safety has intensified alongside the rapid adoption of electric vehicles, empirical and data-driven research remains limited in the South Korean context. To address this gap, the study employs a suite of multivariate statistical techniques-including descriptive statistics, chi-square tests, binary logistic regression, and hierarchical cluster analysis-to examine fire occurrence patterns and risk profiles. The analysis reveals statistically significant associations among battery manufacturer, operational status at the time of ignition, and the initial ignition source. By statistically validating domestic EV fire data, this study provides evidence-based insights into fire causation mechanisms and proposes targeted technical improvements and policy recommendations to enhance electric vehicle safety.

This study was conducted in response to the CCPI 2025 report, which ranked South Korea among the lowest countries in greenhouse gas reduction performance, underscoring the need for more robust environmental policies. In particular, research efforts have intensified to address the slowdown in emission reduction within the transportation sector. This study analyzes the structural safety degradation of the main frame caused by the heavy battery used in EV-converted vehicles and proposes an improved subframe mounting method to extend the operational lifespan of such converted vehicles. Although conventional L-shaped brackets offer easy installation and low cost, they pose potential safety risks during long-term operation after conversion. Therefore, a refined subframe cross-sectional design was developed to distribute loads concentrated on the main frame, thereby reducing fatigue deformation and structural failure. The installed subframe was analyzed under three joint conditions-welded, bolted, and hybrid (weld-bolt combined)-to evaluate structural safety. Based on the comparative structural analysis results, the study presents an optimized subframe mounting method suitable for EV-converted light trucks.

This study develops an SOH prediction model for electric vehicle batteries using periodic technical inspection data collected under driving conditions. As the EV market expands rapidly, accurate SOH estimation is essential for battery safety management, maintenance optimization, and circular economy policy design. The model employs Ridge regression with L2 regularization to address multicollinearity among key predictors as vehicle age, cumulative mileage, total operating time, and total discharge. Cross-validation results showed MAE=2.24%p and RMSE=2.74%p, outperforming a random forest baseline. Variable importance analysis revealed that age and mileage were the strongest predictors, with calendar aging and cycle aging contributing comparably to SOH degradation. A framework is proposed to estimate time-to-threshold for SOH criteria using daily usage patterns. This supports targeted inspection scheduling, residual value assessment for used EVs, and policy development for battery reuse and recycling. By leveraging large-scale field inspection data, the research bridges the gap between laboratory-based SOH modeling and practical diagnostic applications for EV safety management.