Crowdsourcing-based traffic simulation for smart freight mobility

Abstract

Crowdsourcing is emerging as a powerful tool in transportation applications as it aims to provides possible solutions to problems by soliciting inputs from the crowd - usually people, objects or entities on an individual level. The primary sources of ‘big data’ in transportation (which includes social media, mobile applications and connected vehicles) facilitate this evolution through crowdsourcing; however, limited applications of crowdsourcing have been found in literature for the freight operations. In this paper, a simulation framework is developed combined with probabilistic modeling to make freight truck a ‘smart freight’ truck - with improved mobility by being able to detour to avoid a downstream congestion on the path. Smart freight trucks have access to crowdsourced data on an imminent congestion on its route and can detour leveraging this privilege. A discrete-time Markov chain (DTMC)-based simulation model is developed which describes the detour process through the nearest exit ramp of a freeway. The detour occurs with the help of crowdsourced information on a downstream congestion location. Simulation analyses carried out with both average daily passenger cars and truck traffic volumes using interstate 405 (I-405) and I-605 in the Southern California Region substantiate the effectiveness of the model. Simulation analysis shows that the efficiency improvements in detouring freight trucks under free flow traffic conditions and congested scenarios are 52% and 31%, respectively. Since connected vehicle technology (CVT) relies heavily on crowdsourced information from other vehicles in the fleet, simulation findings of this research can have large-scale implications in determining the success of CVT applications in freight operations. This crowdsourced data-based simulation framework can be used within traffic simulation software packages to enhance both passenger and truck freight operations.