Journal of Transportation Engineering, Part A: Systems

The durability of bridge deck pavement significantly influences the economic gains from bridge maintenance and transportation efficiency. With the increasing application of autonomous trucks (ATs) in highway bridges, there is a growing need to ascertain the adequacy of existing bridge deck pavement designed for human-driven vehicles to withstand the unique driving behaviors of ATs, especially their distinctive lateral driving behaviors, such as centering within lanes, which may affect the lifetime of bridge deck pavement. In this study, we first establish both the mechanical model and continuous variable temperature field model of the bridge deck pavement. Subsequently, a durability evaluation method is proposed for assessing rutting in bridge deck pavement that accounts for the lateral driving behavior of ATs. We examine the impact of varying lateral driving behaviors of ATs within a lane on the durability of bridge deck pavement with a double-layer structure. The findings show that the regulation of the lateral driving behavior of ATs can effectively postpone the onset of bridge deck pavement deterioration akin to rutting in human-driven trucks by up to 3.66 years (as determined by periodic detection year). This results in a notable reduction in rutting of 11% compared to human-driven trucks. Conversely, without the implementation of lateral control, the rutting caused by ATs could increase by 54%. Moreover, we develop and formulate lateral control strategies for ATs based on the theoretical insights obtained. These strategies take into consideration the applicable conditions for effective lateral control. This investigation not only provides essential theoretical support for maintenance and design strategies of bridge deck pavement within the realm of ATs but also offers operational recommendations for ATs on highway bridges. These suggestions encompass traffic speed limits and prescribed driving behaviors within lanes, which are essential considerations from the standpoint of AT operations.

To facilitate cooperation among connected automated vehicles (CAVs), such as cooperative platooning or cooperative lane changes in mixed traffic comprising CAVs, connected human-driven vehicles (CHVs), and conventional (unconnected) human-driven vehicles (HDVs), an ego CAV needs to discern which surrounding vehicles are connected (CAVs or CHVs) and which are unconnected (HDVs). Therefore, this paper introduces a surrounding vehicles identification system (SVIS) designed to enable CAVs to identify the connectivity status of surrounding vehicles. Unlike a previously developed preceding vehicle identification system that only identifies the connectivity of its immediately preceding vehicle, the proposed SVIS generalizes this identification to nearby surrounding vehicles. Furthermore, the proposed SVIS significantly reduces the error in the connected vehicle identification by integrating distance and the magnitude of velocity for matching instead of distance-only matching used by the previous system. The SVIS is a must to enable CAVs to form platoons via lane changes after a connected vehicle is detected in an adjacent lane. We compared a newly proposed distance-plus-speed-matching approach with a previously developed distance-only-matching approach using Next Generation Simulation (NGSIM) data. Through comparison, the efficacy of the proposed distance-plus-speed-matching SVIS is demonstrated to be considerably superior to that of the distance-only-matching SVIS. Given that a reliable and robust method for identifying surrounding connected vehicles is key to the successful formation of cooperative adaptive cruise control (CACC) platoons and cooperative lane changes, the proposed distance-plus-speed-matching SVIS would help quickly and efficiently form platoons in a mixed traffic environment.

Abstract

Turbo roundabouts, originating from the Netherlands, have been proven to have safety and operations benefits compared to traditional multilane roundabouts. Turbo roundabouts are gradually being adopted by North American transportation and road authorities. However, no design guidelines have been developed yet per North American practices. Sight distance requirements were not even specifically addressed in the Dutch mature design guidelines for turbo roundabouts. This paper focuses on basic turbo roundabouts, examines stopping sight distance (SSD) characteristics, recommends SSD requirements for three critical locations—circulatory roadway, entry approach, and crosswalk on the exit approach—based on AASHTO and Transportation Association of Canada’s (TAC’s) SSD guidelines, presents analytical models, and develops design guidelines for the required lateral clearance and intersection clear sight triangles to meet the required SSD. The values of the required SSD at the circulatory roadway and to the crosswalk on the exit approach equal the values for the speeds. The required approach SSD is the same as for any roadway section with the same design speed. The minimum crosswalk setback of a one-vehicle gap of 6.0 m between the queued vehicle and the crosswalk is sufficient to meet the required SSD. The required lateral clearance and intersection clear sight triangles are recommended. The recommended guidelines would help develop future design guides per North American practices and provide a guide to property needs for planning basic turbo roundabouts.

Practical Applications

Turbo roundabouts initially built in the Netherlands have gradually attracted North American transportation and road authorities’ attention, given their safety and operations benefits. After more than a decade’s effort in advancing and promoting turbo roundabouts, the first turbo roundabout was built in the US in 2022. However, design guidelines per North American practices are lacking, such as the required stopping sight distance (SSD). The required SSD at the circulatory roadway and entry approach and SSD to the crosswalk on the exit approach are three critical locations. The values of the required SSD at the circulatory roadway and to the crosswalk on the exit approach equal the values of the travel speeds. The required approach SSD is no different from any roadway section with the same design speed. The minimum crosswalk setback of a one-vehicle gap of 6.0 m between the queued vehicle and the crosswalk is sufficient to meet the required SSD. Sight obstructions should meet the required lateral clearance at the roundabout approach and circulatory roadway due to the curve alignment. Clear sight triangles at the roundabout quadrants should be provided to meet the required SSD to the crosswalk on the exit approach.

Microscopic modeling of vehicle movements and interactions is pivotal in traffic flow theory. Physics-based car-following (CF) models using mathematical formulations can delineate driving behavior in various traffic conditions with decent interpretability. However, given predetermined mathematical forms, they might fail to characterize complex, highly nonlinear phenomena. Data-driven CF models naturally excel in this regard considering their flexible architectures, but their performance is subject to data quality, especially distribution bias. In this paper, we propose a novel physics-informed particle filter (PIPF) model that fuses and takes advantage of the two approaches. Utilizing the intelligent driver model as the physics-based model and the multioutput Gaussian process regression as the data-driven model, the PIPF model integrates and embeds both models into a particle filter framework, enhancing both model adaptability and accuracy. The performance of the proposed model is examined through both single vehicle and multivehicle numerical experiments using the NGSIM trajectory data set. Compared with physics-based and data-driven models alone, the PIPF model demonstrates a performance improvement in terms of the root mean square error of about 11.16% and 29.43% in scenarios characterized by sparse data and about 19.81% and 3.84% in scenarios with sufficient data. Compared to traditional particle filtering models, the number of particles to achieve optimal results is reduced by 20%, meaning less computational complexity.