Developing A Car-Following Model With Consideration Of Driver'S Behavior Based On An Adaptive Neuro-Fuzzy Inference System

Junhua Wang,Lanfang Zhang,Siwen Lu,Zhongren Wang
DOI: https://doi.org/10.3233/IFS-151770
2016-01-01
Journal of Intelligent & Fuzzy Systems
Abstract:Based on a car following experiment with a synchronized fifth-wheel system, the Adaptive Neuro-Fuzzy Inference System (ANFIS) is utilized to develop a car following model. Wavelet analysis was used to denoise the high frequently collected fluctuating data. Drivers' reaction time was calculated using Cross Correlation analysis. In the car-following model developed, relative speed, distance headway between leading and following vehicle, and speed of following vehicle are the three input parameters; while acceleration of the following vehicle is the output. Comparing with the traditional General Motor (GM) car-following model, the results show that the ANFIS-based car-following model can better describe the freeway driving behavior.
What problem does this paper attempt to address?