This paper proposes a scalable and interpretable framework for lane-wise highway traffic anomaly detection, leveraging multi-modal time series data extracted from surveillance cameras. Unlike traditional sensor-dependent methods, our approach uses AI-powered vision models to extract lane-specific features, including vehicle count, occupancy, and truck percentage, without relying on costly hardware or complex road modeling. We introduce a novel dataset containing 73,139 lane-wise samples, annotated with four classes of expert-validated anomalies: three traffic-related anomalies (lane blockage and recovery, foreign object intrusion, and sustained congestion) and one sensor-related anomaly (camera angle shift). Our multi-branch detection system integrates deep learning, rule-based logic, and machine learning to improve robustness and precision. Extensive experiments demonstrate that our framework outperforms state-of-the-art methods in precision, recall, and F1-score, providing a cost-effective and scalable solution for real-world intelligent transportation systems.
Ensuring the safety of vulnerable road users (VRUs), including pedestrians, cyclists, electric scooter riders, and motorcyclists, remains a major challenge for advanced driver assistance systems (ADAS) and connected and automated vehicles (CAV) technologies. Real-world VRU tests are expensive and sometimes cannot capture or repeat rare and hazardous events. In this paper, we present a lightweight, configurable simulation framework that follows European New Car Assessment Program (Euro NCAP) VRU testing protocols. A rule-based finite-state machine (FSM) is developed as a motion planner to provide vehicle automation during the VRU interaction. We also integrate ego-vehicle perception and idealized Vehicle-to-Everything (V2X) awareness to demonstrate safety margins in different scenarios. This work provides an extensible platform for rapid and repeatable VRU safety validation, paving the way for broader case-study deployment in diverse, user-defined settings, which will be essential for building a more VRU-friendly and sustainable intelligent transportation system.
Vehicular communication allows vehicles to interact with road users, roadside infrastructure, and cloud-connected devices. It holds a crucial position in modern transportation systems, impacting both fundamental and advanced aspects and enhancing traffic safety and efficiency. C-V2X is a wireless communication technology that uses cellular networks to enable communication between vehicles and infrastructure. C-V2X can be used for applications such as collision avoidance, traffic management, and remote vehicle diagnostics. This project conducted a feasibility study on the current position of C-V2X in the industry and developed a prototype, RampCast, to fundamentally understand the current C-V2X implementations as part of the 3GPP Release 14. A comprehensive review on the state-of-the-art CV2X technologies and various demonstration projects were carried out by the automotive industry, cellular wireless chips/systems companies, and federal/states DOTs in the U.S. and Europe. A geocast-based prototype system, named RampCast, was built using a software-defined radio approach. The RampCast algorithms focused on the geocasting and were developed for improving message prioritization and retransmission. The field tests that were conducted in a campus parking lot and on the test track revealed sub-100 ms latency and a range of up to 2,500 ft for C-V2X, which emphasized its effectiveness in transmitting critical messages and traffic guidance. Further extensions for the prototype include incorporating multiple units, expanding message types (e.g., points of interest and location-specific adverts), optimizing the prototype's GUI for diverse scenarios, and conducting long-term data analysis for better message flow optimization.