Examining the Limitations for Visual Angle-car following Models

Visual angle car following models assume fixed values (thresholds) for the angular velocity (∆Ө/∆t) over which the driver of the following vehicle is affected by the leading vehicle. This paper explains the advantages and limitations of such models and provides possible solutions for these limitations. Sensitivity analysis has been carried out for these purposes as a part of this paper. It was found that the main advantages of using visual angle models are their simplicity and ability to represent the effect of the widths of different types of vehicles (e.g. heavy goods vehicles, cars and motorbikes) on space headway (i.e. the clear distance between leading and following vehicles). Also, these models are simply capable of including the effect of driver’s reaction time as a function of traffic density, relative speeds between leading and following vehicles and the action/behaviour of the leading vehicle (i.e. accelerating/decelerating following a change in local traffic conditions). However, the main concern in using such models is related to the appropriate choice of the threshold values. The paper will illustrate that using a fixed threshold value for the angular velocity will be illogical since it eliminates the effect of variations in speed of the following vehicle for the same given relative speed between the leader and follower (∆V). Moreover, other limitations are related to the effect of leader’s width on the acceleration/deceleration rates of the follower. Finally, and in order to deal with the above limitations and advantages that visual angle models have, an alternative car following model has been proposed. The proposed model will then be tested against data from sites to examine its applicability.