01152559

Component

In past years, a series of models have been built up to describe driving behaviors in a traffic stream. Fuzzy logic is one of the ways that can simulate and model the driver's car-following behavior. This research aims at using a two-order fuzzy table look-up scheme to develop a microscopic car-following model, and to simulate driver's acceleration behaviors from detailed vehicle trajectory data. Driver specialty, which can be derived by the first order table look-up scheme, is introduced to analyze various behaviors of different drivers. The modeling section begins with using the fuzzy c mean to build up membership functions. The fuzzy rules are selected with the maximum degree from the rule base and the missing cells in the rule table are filled by the initialed advanced K-nearest neighbor (KNN) algorithm. A modeling process is built up using a self-developed MATLAB program, and the simulation is conducted and the comparison between the real trajectory data and simulation data is conducted. The results have shown the feasibility of using the designed model to simulate the acceleration behavior.

01147878

10-2972

English

21p

2010

Transportation Research Board 89th Annual Meeting

DVD

Figures (7) ; References (18) ; Tables (1)

Acceleration (Mechanics); Car following; Drivers; Fuzzy algorithms; Fuzzy logic; Operating speed; Tables (Data); Traffic simulation

K-Nearest Neighbor Models

Data and Information Technology; Highways; I71: Traffic Theory

Transportation Research Board Annual Meeting 2010 Paper #10-2972

BTRIS, TRIS, TRB

Jan 25 2010 11:27AM