基于GPS定位数据的高速公路换道特征分析与行为识别

Analysis and recognition of highway lane-changing behavior characteristics based on GPS location data

为了实现对高速公路微观状态的监测,需要对车辆换道行为进行识别.利用车载GPS定位数据获取车辆换道参数,并分析换道车及其周围车辆在换道时的参数统计分布特征,提取换道行为表征参数,建立了基于隐马尔可夫的换道行为识别模型.结果表明:利用该模型可以实现对驶出换道、驶回换道和车道保持行为的较好识别,且能在换道发生后1s识别出换道行为,准确率高达92%.为后期实时换道预警提供参考.

In order to realize the identification and monitoring of the microscopic state of highway traffic flow, it is necessary to identify and monitor the lane-changing behaviors of the vehicles. In this paper, we acquire the lane-changing parameters by exploiting on-board GPS（Global Positio- ning System） data. Then, the parameters＇ statistical distribution characteristics of vehicles that change lanes and their surrounding vehicles are analyzed and the characterization parameters of lane-changing behaviors are extracted at the same time, and on those grounds the lane-changing i- dentification model is established based on Hidden Markov Model （HMM）. The results show that the proposed model can accurately identify the lane-changing behaviors, including outbound lane-changing behaviors, inbound lane-changing behaviors and lane keeping behaviors. Moreover, these behaviors can be identified 1 s after they really happen with 92% accuracy rate. It can pro- vide reference for real-time lane change early-warning in future.