Modelling of driver's steering behaviour control in emergency collision avoidance by using focused time delay neural network

Abstract

This paper presents a modelling approach of human driving behavior in emergency rear-end collision avoidance focusing on steering maneuver. The target scenario is set up under real experimental environment and the naturalistic data from the experiment are collected. Dynamic Artificial Neural Network which is Focused Time Delay Neural Network (FTDNN) is used to model drivers steering behaviour. From the obtain results, it can be concluded that the FTDNN model able to simulate drivers steering maneuver in rear-end collision avoidance with the accuracy of which the coefficient determination is 99% (0.99). With further study, this model would beneficial to design motion control strategy to improve Advance Driver Assistance System (ADAS) in collision avoidance system.