Investigation of driving behaviour transition during evacuation and its implication for traffic flow operations based on an open-source traffic simulation platform

Earlier studies have shown how driving behaviour during emergency evacuation is different from that under “normal” conditions. At the same time, most evacuation studies (where the driving performance is explicitly accounted for) were conducted under the assumption of a fully activated emergency environment. However, the transition phase in evacuation, from normal driving situations to emergency driving situations, and also its implication for traffic management has not yet been widely studied. During this phase, drivers receive evacuation information and instructions at different levels, which may induce heterogeneous behavioural driving response. The action and interaction of heterogeneous drivers would influence traffic flow operations. This paper is the pioneer to investigate this transition phase at the early stage of emergency situations and its implications for traffic flow operations. This is tested via a simulation study using a recently-developed open-source traffic simulation platform, where the driver behaviour (activation level, transition period) can be specified. Various scenarios are modelled showing the effect of time-varying (heterogeneous) individual driving behaviour based on a set of performance indicators at both network and individual levels (network clearance time, arrival rate, travel time). The analyses show that high penetration of activated drivers with a high activation level and a short transition period can significantly improve network performance. Therefore, traffic control and management can influence driving behaviour of evacuees towards such an optimal operation regime to improve evacuation efficiency.