Vehicle trajectory reconstruction using automatic vehicle identification and traffic count data

Abstract

SUMMARYThe origin-destination (OD) matrix and the vehicle trajectory data are critical to transportation planning, design, and operation management. On the basis of the deployment of automatic vehicle identification (AVI) technology in urban traffic networks in China, this study proposed a vehicle trajectory reconstruction method for a large-scale network by using AVI and traditional detector data. Particle filter theory was employed as the framework for this method that combines five spatial-temporal trajectory correction factors (i.e., the path consistency, the AVI measurability criterion, the travel time consistency, the gravity flow model, and the path-link flow matching model) to estimate the trajectory of a vehicle. The probabilities of the most likely trajectories are updated by the Bayesian method to approximate the 'true' trajectory. The traffic network in the Beijing Olympic Park was selected as the test bed and was simulated by using VISSIM to create a complete set of vehicle trajectories. The accuracy of the resulting trajectory reconstruction exceeds 90% when the AVI coverage is only 50%, assuming an AVI detection error of 5% for a closed network and an open network. The average relative error of a static OD matrix is 11.05%. Although the accuracy of reconstruction exceeds 80% when the AVI coverage is between 50% and 40%, the accuracy of a defective product-OD matrix decreases rapidly. The proposed method yields high estimation accuracy for the full trajectories of individual vehicles and the OD matrix, which demonstrates significant potential for traffic-related applications.

Outlier detection from vehicle trajectories to discover roaming events

“…Generally, real data are noisier than synthetic data and should be preprocessed with data cleaning methods as indicated in typical data mining processes. Thus, additional trajectory reconstruction [13,38] or probabilistic prediction [4,34] approaches may be exploited to complement our approach. Third, POI mentioned in Section 4.1 might reduce the number of false positives.…”

Vehicle path reconstruction using automatic vehicle identification data: An integrated particle filter and path flow estimator

“…The implementation of PF is typically challenging in the vehicle path reconstruction problem based on AVI data alone, because limited information can be directly obtained from AVI data. Regardless of the usage of the loop detector data or AVI data, two observation models (i.e., the path consistency model and the AVI measurability criterion model), which are detailed in Feng et al (2015), are equivalent. Two additional observation models are expressed as follows:…”

“…Read the PFE estimated link travel times t a from Step 1. Generate the PF possible path set K possible for all OD pairs using the depth-first search algorithm adopted in Feng et al (2015), which is based on estimated travel time and network topology.…”

“…An AVI system can collect information such as vehicle IDs, vehicle passing times and vehicle positions. In previous studies, the collected data from AVI system can be used for origin-destination (OD) matrix estimation (Van Der Zijpp, 1997;Dixon and Rilett, 2002;Zhou and Mahmassani, 2006;Parry and Hazelton, 2012), path flow estimation (Bell et al, 1997;Castillo et al, 2008a;Castillo et al, 2010;Teknomo and Fernandez, 2012), travel time estimation (Dion and Rakha, 2006;Tam and Lam, 2011), traffic flow operation risk assessment (Abdel-Aty et al, 2012;Ahmed and Abdel-Aty, 2013) and vehicle path reconstruction (Feng et al, 2015). These are fundamental issues for urban traffic management.…”

“…The PF can provide a visual representation of the similarity between a partial path and the final path, which can be used to directly capture drivers' random perceptions at microscopic level. However, the PF highly dependent on observation models; few observation models can be used when only a set of link flows is known (i.e., two observation models in Feng et al (2015) cannot be directly applied based on AVI data alone). In addition, the PF does not consider the fact that the available counts are inconsistent due to AVI measurement errors.…”