Detection and pre-warning of vehicle lane change based on state machine

Abstract

This paper focuses on the problem of detecting driver's changing lane, and gives a state machine model to analyze driving behavior. Firstly, three-axis accelerated velocity sensor is used to gather data, and then smoothing filtering method is made to process data. Secondly, a change rate of accelerated velocity is computed by using linear regression method, and vehicle's turning range and acceleration\deceleration are analyzed. Thirdly, vehicle's running state is fuzzily processed and a state machine is built to detect vehicle running. Experimental results show that the state model could correctly detect driver's driving behavior and send warning signals.