An adaptive framework for real-time freeway traffic estimation in the presence of CAVs. (April 2023)

• Record Type: Journal Article
• Authors: Makridis, Michail A.; Kouvelas, Anastasios
• Is Part Of: Transportation research. Volume 149(2023)
• Publication Date: 2023-04
• Subjects: Online freeway traffic estimation -- Data assimilation -- Data fusion -- Adaptive unscented Kalman Filter -- Connected and automated vehicles -- Intelligent transportation systems; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2023.104066
• Languages: English
• ISSNs: 0968-090X

Abstract

Abstract: Advancements in sensor technologies, vehicle automation, communication, and intelligent transportation systems create unforeseen possibilities for the development of novel traffic management approaches in road transport systems. Furthermore, data observations with different accuracy and noise levels are fused towards advanced traffic state estimators. This work builds on the existing family of data assimilation techniques in the literature and proposes an online adaptive framework, fusing observations from static and moving sensors, along with estimations inferred from a traffic flow model and performing real-time traffic estimation in the presence of Connected and Automated Vehicles (CAVs). A real-world case study was used for validation and assessment of the proposed framework against well-known methodologies in the literature. The benefits and downsides of each approach for different scenarios are discussed, as well as the performance of each framework for different traffic models and penetration rates of CAVs. Highlights: Propose an Adaptive Unscented Kalman Filter for traffic state estimation. Dynamic update of noise error covariance and readjustment of sigma points. Compare results for AUKF, EnKF, EKF and UKF frameworks for TSE. The role of reporting instantaneous speeds from moving vehicles for TSE. Comparing CTM and METANET performance for TSE on a large case study.