Modelling behavior of cyclists to evaluate bicycle traffic performance

Abstract: Over the past years, bicycles started to gain popularity and more people are now using it as amean of transportation. Nowadays, sustainability trends promote cycling as one of the greenestsolutions for the future in mobility. With this rising demand in the number of bicycle trips, there isa need to design adequate infrastructure fitted to the behavior of cyclists and their traffic patterns.Even though there is plenty information regarding car driving behavior, there is still a lack ofunderstanding in how cyclists behave. Consequently, the behavior of cyclists is often modelled byadjusting parameters from models originally designed for other types of users (e.g. car-followingmodels).The purpose of this degree project is to increase the understanding of the behavior of cyclists byinvestigating the extent to what car-following models can be used to simulate bicycle traffic andwhat effects relevant parameters have on the behavior of simulated cyclists and specific trafficcharacteristics. The main method is microscopic traffic simulation, through the simulation programVissim, of bicycle traffic flow on a two-way separated cycle path in Stockholm, withoutdisturbances from pedestrian and car traffic. In addition, a sensitivity analysis is presented, withthe support of the Elementary Effects method, of behavioral parameters of following, overtaking,lateral position, and the longitudinal gradient of the cycle path, and their effects on speed, lateralposition, and time headway.The sensitivity analysis showed that the minimum lateral distance between cyclists is a keyparameter to model following and overtaking behavior, while the gradient was found to be animportant factor that has a great impact on the bicycle traffic performance. Overall, the resultsindicated that the Vissim implementation of the Wiedemann car-following model is able toreproduce the speed and time headway on flat cycle paths with high flows, with goodapproximation to reality. However, the model failed to represent the effect of gradient, and thelateral position of cyclists could only be achieved with a low level of detail and accuracy.The understanding of the behavior of cyclists is crucial for the design and planning of infrastructurededicated to bicycles, so transport planners can consider the needs of cyclists. This projectcontributes to establish a baseline for bicycle traffic modelling and its adaptation of car-followingmodels to simulate cyclists and bicycle traffic. The awareness of these adaptations will help topromote cycling in order to achieve a more sustainable transport system