A new car following model from the perspective of visual imaging

The paper proposes a car following model from the perspective of visual imaging (VIM), where the visual imaging size of the preceding vehicle on a driver's retina is considered as the stimuli and determines the driving behaviors. NGSIM trajectory data are applied to calibrate and validate the VIM under two scenarios, i.e. following the car and following the truck, whose fitting performance outperforms that of visual angle car following model (VAM). Through linear stability analyses for VIM, it can be drawn that the asymmetry in traffic flow is preserved; the larger vehicle width, vehicle length and vehicle apparent size all benefit enlarging the traffic flow stable region; the traffic flow unstable region when following the car tends to fall in the relatively small distance headway range compared with that when following the truck. After that, numerical experiments demonstrate that the visual imaging information applied in VIM is more contributive to the traffic flow stability than the visual angle information in VAM when following the truck in the relatively large distance headway or involving the driver's perception threshold, i.e. Weber ratio; introducing Weber ratio would break the originally stable traffic flow or deteriorate the traffic fluctuation, which however can be alleviated by increasing drivers' sensitivity, e.g., decreasing Weber ratio. Finally, VIM is verified to be able to satisfy the consistency criteria well from the theoretical aspect.