Predicting the Future State of a Vehicle in a Stop&Go Behavior Based on ANFIS Models Design Ali Ghaffari Mechanical Engineering Department K. N. Toosi University of Technology Tehran, Iran ghaffari@kntu.ac.ir Fatemeh Alimardanii Mechatronics Engineering Department K. N. Toosi University of Technology Tehran, Iran f.alimardani@ee.kntu.ac.ir Alireza Khodayari Mechanical Engineering Department K. N. Toosi University of Technology Tehran, Iran khodayari@ieee.org Abstract— Stop&go cruise system is an extension to ACC which is able to automatically accelerate and decelerate the vehicle in city traffic. There have been attempts to model stop&go waves via microscopic and macroscopic traffic models. But predicting the future state of the maneuver has not attracted much attention. The purpose of this study is to design adaptive neuro- fuzzy inference system (ANFIS) models to simulate and predict the future state of the stop&go maneuver in real traffic flow for different steps ahead. These models are designed based on the real traffic data and model the acceleration of the vehicle which performs a stop&go maneuver. Using the field data, the performance of the presented models is validated and compared with the real traffic datasets. The results show very close compatibility between the model outputs and maneuvers in real traffic flow. Keywords: Intelligent Automation; Stop&go maneuver; neuro- fuzzy inference system; modeling. I. INTRODUCTION Intelligent Transportation Systems (ITS) are being developed and deployed to improve the efficiency, productivity, and safety of existing transportation facilities and to alleviate the impact of transportation on the environment. These systems exploit currently available and emerging computer, communication, and vehicle-sensing technologies to monitor, manage, and control the highway transportation system. The success of ITS deployment depends on the availability of advanced traffic analysis tools to predict network conditions and to analyze network performance in the planning and operational stages. Many ITS sub-systems are heavily dependent on the availability of timely and accurate wide-area estimates of prevailing and emerging traffic conditions. Therefore, there is a strong need for a Traffic Estimation and Prediction System (TrEPS) to meet the information requirements of these subsystems and to aid in the evaluation of ITS traffic management and information strategies [1]. Microscopic models are increasingly being used by transportation experts to evaluate the applications of new ITS. A variety of applications including car navigation systems, adaptive cruise control systems, lanes keeping assistance systems and collision prevention systems directly use the microscopic traffic flow models [2]. To develop a microscopic traffic simulation of high fidelity, researchers are often interested in imitating human’s real driving behavior at a tactical level. In other words, without describing the detailed driver’s actions, driver-vehicle units (DVUs) in the simulation are modeled to replicate their states in reality, i.e., the profiles of vehicle position, velocity, acceleration, and steering angle. Fig. 1 indicates the model structure of a DVU, in which the detailed driver’s actions become internal. A number of factors have been found to influence car-following behavior, and these include individual differences of age, gender, and risk-taking behavior [3]. Figure 1. Structure of a DVU model [3]. Humans play an essential role in the operation and control of human–machine systems such as driving a car. Modeling driver behavior has transferred human skills to intelligent systems, e.g., the adaptive cruise control (ACC) system, intelligent speed adaption (ISA) system, and autonomous vehicles. Human driving models are also indispensable for the performance evaluation of transportation systems. With