This paper discusses the development of a car following model using reactive agent-based techniques based on a neural network approach for mapping perceptions to actions. The model's formulation is similar to the desired spacing models which do not consider reaction time or attempt to explain the behavioural aspects of car following. The models were developed using field car following data, comprising relative speed and headway between vehicles. A number of error tests were used to compare the performance of the agent-based model against established car following models using the field data. The results showed that simple back-propagation neural network models outperformed the Gipps and Psychophysical family of car following models. Qualitative drift behaviour; speed and headway profiles analyses also confirmed the findings. (a) For the covering entry of this conference, please see ITRD abstract no. E211903.
Samenvatting