13 May 2024 in IEEE Sensors Journal
Accurate traffic state estimation, vital for managing urban congestion, is often achieved through simulation. Real-time data is invaluable for this, yet obtaining multi-sensor data is challenging and costly. To bridge this gap, leveraging crowdsourced data from third-party sources, despite its anonymity, enriches available information for precise estimation. This passive crowdsourced data is often reported as estimated time of arrival (ETA) and color-coded traffic patterns. The paper introduces an effective calibration approach for a mesoscopic traffic simulation of a complex urban arterial network, primarily relying on crowdsourced data and incorporating sensor data, if accessible. The approach employs a nested genetic algorithm with average speed data, calculated using ETA to estimate vehicle counts, eliminating the need for time-consuming field surveys. A custom mutation operator-speed deviation adaptive gene mutator, is introduced for generating vehicle counts to replicate real-world traffic conditions in the intermediate time steps of simulation. Additionally, a route selection algorithm is developed using color-coded tracks from Google Maps for prioritizing routes based on congestion patterns. The study demonstrates a novel data-fusion technique, combining sensors with passive crowdsourced information for accurate traffic speed estimation. The proposed methodology was applied to two case studies of urban arterial networks. The values obtained from simulations during validation showed promising proximity to real values, yielding a mean absolute percentage error (MAPE) of 0.29% for a simpler network and 6.31% for the best 10 routes, and 10.33% for all of the 15 priority routes within a complex network.
24 March 2026 in ISPRS Journal of Photogrammetry and Remote Sensing
10 December 2025 in IEEE Transactions on Intelligent Transportation Systems
Vehicular accidents pose a substantial risk to drivers, underscoring the persistent and vital need for heightening safety measures. Early accident anticipation mechanisms are imperative for proactive measures, while detection accuracy is pivotal for prompt response and effective post-accident mitigation. Accurate and early anticipation of accidents for automated driving assistance systems in vehicles or CCTV in cities remains a complex task due to the intricate spatial-temporal interactions within traffic videos. This study presents text-informed magnitude enhancement in contrastive multiple-instance feature learning for vehicle accident detection and anticipation (TIME-VAD). Text is a better representative of concepts when compared to images in video, thus multi-modal learning is suitable. Also, the traditional assumption about feature magnitude of accidents and normal frames in magnitude based multiple-instance learning using weak supervision may not hold. This has led to the development of a novel weak-supervised learning strategy involving magnitude enhancement from textual concepts. For a better frame-level perception of accident risks in videos, dynamic temporal attentions are refined using the proposed dilated temporal conv-attention (DTCA) block. In-depth component-level analysis is performed to showcase the model’s efficacy while elucidating its operational mechanisms. Evaluation is conducted on three benchmark datasets, considering both earliness and accuracy-related metrics. Extensive experiments demonstrate that our TIME-VAD model outperforms the existing models. Compared to the previous top-performing supervised model achieving 84.7% accuracy, TIME-VAD achieves a 94.44% accuracy (measured by ROC-AUC) on the largest DoTA dataset. Notably, our model also excels in measuring how early it detects accidents compared to previous methods. The code will be released on https://github.com/sumitmishra209/TIME-VAD
1 January 2026 in IEEE Transactions on Intelligent Transportation Systems