Lane-Changing Behavior Recognition At The Expressway Exit Based On Power Spectrum

The meaning of vehicle bypass roads refer to expressway exit roads and a section of gradual change roads on highway deceleration lane to shunt side.If vehicles want to shunt,they must need to change lane.So,not only the transverse and longitudinal movement of the vehicle,but also the kinematic relationship between the vehicle and other vehicles in adjacent lanes are involved during lane change operation.Accidents happen frequently at the expressway exit because of the complex environment.In order to achieve the greater security when the vehicle change lane at the exit of expressway,this paper establishes safe and dangerous change lane driving behavior database respectively by collecting change lane information of vehicles at the exit of expressway.The lane changing driving behavior to be tested is matched with the power spectrum of safe and dangerous lane changing behavior to judge the safety status of the vehicle.In this way,the recognition rate of the safety status of the lane changing behavior can reach 98.3%.The research content includes the following points:First of all,based on the real lane-changing behavior data of lane,some characteristic parameters such as the speed of the vehicle,acceleration of the vehicle,displacement are extracted.The collision time TTC of the vehicle is calculated,which can be used to identify risk rank.According to the threshold value of TTC,the lane change driving behavior data can be divided into the safety data and danger data.Second,through analyzing four kinds of lane-changing scene,including lane-changing vehicle with the original vehicle lane,next cars and target lanes,the next car change lane,then according to the laws of the lane changing kinematics combined with vehicle driver’s brain wave change in the process of lane changing to set up lane-changing model of driving.Then,the sequence x(n)formed by sampling the channel change model is used for classical power spectrum estimation and modern power spectrum estimation.We find that x(n)is more suitable for modern power spectrum estimation.In addition,the AR parameter model is analyze The method of order selection is estimated through analysing AR model spectrum estimation method.so that channel changing behavior spectrum is established.Finally,Characteristic parameters matrix about lane changing driving behavior is established,the safe and dangerous libraries of lane changing driving behavior are built up.The power spectrum of a single characteristic of lane changing driving behavior,such as relative distance,relative speed and relative acceleration,is matched with the corresponding safety and danger libraries to obtain the power spectrum matching value.At the same time,the power spectrum of lane changing driving behavior based on the lane changing model is matched with the corresponding safety and danger libraries to obtain the power spectrum matching value.Through judging the safety of driving behavior of lane using the power spectrum matching value,the power spectrum matching accuracy of lane changing model is 98.47%.The power spectrum matching based on the lane changing driving behavior model at the expressway exit is more accurate than the power spectrum matching based on the single characteristic value of lane changing behavior,which also proves the feasibility of the power spectrum matching based on the lane changing driving behavior model to identify the dangerous lane changing behavior.