An increasingly emphasized research area is the forecast of short-term traffic conditions for nonrecurring traffic dynamics caused by random highway incidents such as crashes or roadway closures. This research proposes a prediction framework which focuses on training a machine learning (ML) model to predict the speed heatmap associated with incidents. Heatmaps contain ideal information that depicts the spatiotemporal characteristics of incident-induced impacts and are suitable objects for ML models to understand and predict. Because of the sparsity of incident data in the real world, we proposed a simulation approach to rapidly expand the training dataset, thus speeding up the model training process. The conditional deep convolutional generative adversarial nets is employed to predict the speed heatmap and the mesoscopic dynamic traffic assignment model DynusT was used to generate many training data. The evaluation shows that the proposed model captures both the tonal and spatial distribution of pixel values at 80.19% similarity between the prediction and actual heatmaps. To the best of our knowledge, this is one of the first attempts in the literature to train ML to predict heatmap representation of incident-induced spatiotemporal impact, and speeding up the training via simulation.

No abstract

The initial hype around Automated Vehicle (AV) technologies has subsided, and it is now being realized that near-term deployment of AV technologies will be in the form of low-speed shared automated shuttles in geofenced districts with a high density of trip demand. A concept labeled ‘Automated Mobility Districts’ (AMD) has been coined to define such deployments. A modeling and simulation toolkit that can act as a decision support tool for early-stage AMD deployments is desired for answering the questions such as (i) for a series of given conditions, such as the amount of travel demand and automated shuttle fleet configuration, what is the expected mode split for shared automated vehicle (SAV) services? (ii) for that mode share of SAVs, what level-of-service and network performance can be anticipated? To answer these research questions, an innovative and integrated framework of multi-mode choice and microscopic traffic simulation model is presented to obtain the equilibrium of mode split for various modes in AMDs, based on real-time traffic simulation data. The proposed framework was tested using travel demand and road network data from Greenville, South Carolina, considering a car, walk, and two SAV on-demand ridesharing modes in a proposed AMD. Results from the study demonstrated the efficacy of the proposed framework for solving the mode split equilibrium in an AMD. In addition, sensitivity analyses were conducted to understand the impact of factors such as waiting times and fleet resources on mode share equilibrium for SAVs.