As the deployment of Automated Driving Systems (ADS) and Advanced Driver Assistance Systems (ADAS-L2) expands, understanding their real-world safety performance becomes essential. This study examines the severity and contributing factors of crashes involving vehicles equipped with ADS and ADAS-L2 technologies using NHTSA data. Using machine learning models on crash datasets from 2021 to 2024, this research identifies patterns and risk factors influencing injury outcomes. After data preprocessing and handling missing values for severity classification, four models were trained: logistic regression, random forest, SVM, and XGBoost. XGBoost outperformed the others for both ADS and ADAS-L2, achieving the highest accuracy and recall. Variable importance analysis showed that for ADS crashes, interactions with other road users and poor lighting were the strongest predictors of injury severity, while for ADAS-L2 crashes, fixed object collisions and low light conditions were most influential. From a policy and engineering perspective, this study highlights the need for standardized crash reporting and improved ADS object detection and pedestrian response. It also emphasizes effective human–machine interface design and driver training for partial automation. Unlike previous research, this study conducts comparative model-based evaluations of both ADS and ADAS-L2 using recent crash reports to inform safety standards and policy frameworks.

As Advanced Driver Assistance Systems (ADAS) and Automated Driving Systems (ADS) technologies become more prevalent, evaluating their safety performance in complex, real-world environments has emerged as a critical research priority. This study presents a systematic literature review of crashes involving vehicles equipped with ADAS and ADS. Using the PRISMA methodology, this review analyzed 51 peer-reviewed articles and books published between 2018 and 2025. Articles were identified through comprehensive searches of ScienceDirect, TRID, and the Transportation Research Record. The selected studies were organized into five thematic categories: safety and crash characteristics, human interaction and user behavior, testing tools and simulations, cybersecurity and perception risks, and policy and governance. Findings reveal that while AV technologies show promise in reducing crash severity and frequency, performance is highly context-dependent, affected by roadway geometry, weather, system limitations, and user behavior. ADAS-related crashes are more common on highways and often involve human error, whereas ADS crashes are more likely in urban settings and linked to system disengagements. The review also highlights significant gaps in data availability, standardized evaluation protocols, and human-machine interaction research. Despite growing literature, many studies rely heavily on simulations and legacy datasets, limiting generalizability. To support safer AV deployment, future research must prioritize real-world testing and user trust frameworks deployment strategies. This review offers a knowledge base to guide policymakers, researchers, and industry stakeholders in advancing the development and regulation of vehicle automation technologies.

Background: Pedestrian fatalities remain a critical urban transportation safety concern, with crash outcomes potentially influenced by temporal conditions, vehicle characteristics, and contributing factors. This study investigates the associations between these characteristics and pedestrian fatal-crash outcomes in New York City (NYC) from 2021 to 2025, with particular attention to nighttime occurrence and solar darkness.Methods: Police-reported collision data from NYC Open Data were analyzed. Of 486,548 collision records, 42,911 pedestrian casualty crashes were identified, including 606 crashes involving at least one pedestrian fatality. Geographic information system (GIS) techniques were used to recover missing borough information. Descriptive statistics, chi-square tests, and multivariable logistic regression models were employed to examine temporal patterns, first-recorded vehicle types, and police-reported contributing factors. Solar darkness was classified using crash-specific sunrise and sunset times for 40,578 crashes. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were estimated across five regression specifications.Results: The fatal-crash proportion was higher during nighttime (2.62%) than daytime (1.03%). In the primary regression model, nighttime was associated with increased odds of pedestrian fatal-crash occurrence (aOR = 2.55, 95% CI: 2.16–3.00), and the association persisted after additional adjustment for vehicle types and police-reported contributing factors (aOR = 2.72, 95% CI: 2.30–3.23). Exploratory analyses identified elevated fatal-crash odds for crashes involving first-recorded heavy/commercial vehicles, buses, pickup trucks, and SUVs/station wagons. Solar darkness was also associated with higher fatal-crash odds (aOR = 2.59, 95% CI: 2.16–3.09), and the association remained significant after adjustment for calendar month and five time-of-day categories (aOR = 2.32, 95% CI: 1.67–3.22). Weekend associations were less consistent across model specifications.Conclusions: Nighttime was consistently associated with increased odds of fatal outcomes among reported pedestrian casualty crashes, while vehicle characteristics and police-reported contributing factors revealed additional differences in crash outcomes. Solar darkness provided complementary information beyond conventional clock-based nighttime definitions. These findings support integrating temporal, vehicle-related, and environmental characteristics into urban pedestrian safety assessments to better identify conditions associated with fatal crashes.