The weights of drivers’ on-route concerns in car-following behaviours

Among the available studies run on reproducing, learning and classifying driving styles, only few have quantified the weights of drivers’ on-route concerns. A new method is proposed to quantify the issue at hand in car-following behaviours, in which a multi-objective optimisation model with crash risk, comfort and travel time objectives as the driver’s decision-making process is trained. This model is trained based on different car-following trajectory real data, distinguished from each other by the driving phase and behaviour of the following vehicle, where the best-matched weights for the model’s objectives are identified. The results indicate that the weights obtained for the crash risk and comfort objectives are of more concern compared to travel time. The weights of the crash risk, comfort and travel time objectives are 0.4, 0.4 and 0.2, respectively. In deceleration driving phases, timid and aggressive drivers weigh the model’s objectives differently.