Onramp merging zones are critical areas in highway networks where interactions between mainline and onramp vehicles often disrupt traffic flow, causing congestions and safety concerns. The transition toward mixed-traffic environments, where human-driven vehicles (HDVs) coexist with connected and automated vehicles, introduces complexity because of behavioral variability and uncertainty in HDV behavior. Irregular HDV behaviors such as sudden lane changes and speed variance, challenge intention prediction. Monitoring and detecting HDV intentions near onramp merging zones is essential to ensure safe and efficient traffic flow. Distributed fiber optic sensing (DFOS) technology offers promising capabilities for real-time, continuous, wide-range vehicle behavior monitoring but faces limitations in detecting key vehicle attributes such as type, size, and lane-of-travel. In this paper, we propose an integrated traffic monitoring methodology that combines high accuracy and reliability of cameras that measure vehicle attributes with wide-area monitoring capabilities of DFOS systems to enable robust vehicle tracking and behavior estimation. The proposed sensor fusion involves detecting and tracking vehicles within the camera field-of-view, matching vehicles with their corresponding DFOS trajectories, and monitoring vehicle behavior by detecting lane changes. This approach enables robust vehicle tracking and behavior estimation by matching camera-derived vehicle attributes with corresponding DFOS-derived vehicle properties. The proposed methodology was validated through a field trial conducted on the Shin-Tomei expressway in Japan. The evaluation demonstrated vehicle matching accuracy of 92% and continuous lane-change detection accuracy of 82%, respectively. These results highlight the potential of combining DFOS and point sensors to support safe and efficient onramp merging in mixed-traffic environments.
Prasad et al. (2026) studied this question.