Drivers' car-following correlative behavior with preceding vehicles in multilane driving

Abstract

Modelling and analysing drivers' behaviour are the foundation to develop advanced driver assistance systems (ADAS). However, no matter what kind of model structure and expression form are applied, the current method of driver characteristics analysis only considers the driving state of preceding vehicle in the host lane, neglecting the driving state of vehicles in the adjacent lane. Based on experimental data from real-road tests, the traditional analysis of host-lane driving characteristics has many limitations. To address the uncertainty of drivers' behaviour, Grey System Theory is introduced, and the described method of car-following behaviour in multilane driving is established to analyse the difference when following a host-lane vehicle versus following an adjacent-lane vehicle. Results show that traditional car-following behaviour analysis is only applicable when the host-lane preceding vehicle has a small headway. When it is long distance, a driver's risk perception is subject to the combined effects of multiple preceding vehicles. Preliminary research into car-following behaviour in multilane situations shows that the result of multilane analysis is not entirely consistent with that of single-lane analysis, which indicates that it is necessary to explore drivers' behaviour rules in multilane driving to improve both adaptability and acceptability of current advanced driver assistance systems (ADAS).