Distributed Microscopic Traffic Simulation with Human-in-the-Loop Enabled by Virtual Reality Technologies

• DOI: 10.1016/j.advengsoft.2021.102985
• Publication Date: 2021-02-23T00:51:10Z
• Authors: Mahmud Hasan; Daniel Perez; Yuzhong Shen; Hong Yang
• Keywords: 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

Abstract Microscopic traffic simulation (MTS) is the emulation of real-world traffic movements in a virtual world with various traffic entities (e.g., vehicles and pedestrians). Typically, the movements of these entities in the simulation are governed by some pre-defined algorithms (e.g., car-following models and lane-changing models). Modelers may develop customized algorithms through an application programming interface (API). In comparison to the human-controlled vehicles, however, any algorithm will fall short. Other than the configuration of some simplified behavioral parameters (e.g., reaction time), most of the existing MTS models have limited human-in-the-loop simulation abilities to capture the interactions among simulated entities. Besides, most existing MTS models do not provide a realistic virtual environment that enables high-fidelity simulations of the driver behavior in response to various driving conditions, such as road and weather conditions. This paper aims to develop a framework for improving MTS models and extending their capabilities by incorporating distributed vehicles controlled by human-in-the-loop with virtual reality (VR) technologies. This framework supports not only multiple/many users from geographically distributed locations but also interactive visualizations with VR devices from the perspective of a driver and other users (e.g., pedestrians). Data collected from the user-controlled vehicles can be used to calibrate different algorithms like car following, lane changing, etc., or to explore the virtual world from different perspectives for design investigation by potential stakeholders. This paper describes the overall framework, the significant challenges in the proposed approach, and our solutions to these challenges.