提出了一种驾驶分心辨识方法,该方法采用的信号可从配备了车道偏离报警系统的车辆上获得。在驾驶模拟器上,驾驶员通过在驾驶过程中执行第二任务模拟分心驾驶.从而获得专注与分心两种驾驶状态下的数据。采用车辆偏航角以及转向盘转速的标准差作为辨识特征量.基于高斯混合隐马尔可夫模型.建立了专注与分心的驾驶员状态模型。对模型的离线验证表明,该方法对驾驶分心辨识具有较高的准确率。
The paper presents a driving distraction recognition method, which uses readily-available signals from a vehicle equipped with lane departure warning system. By performing the secondary task to simulate distracted driving, data of concentrated driving and distracted driving is obtained on the driving simulator. The standard deviation of vehicle yaw angle and steering wheel velocity are used as recognition characteristic value. Concentrated driving state model and distracted driving state model are established based on Gaussian Mixture-Hidden Markov Model (GM-HMM). The model offline proves that this method can accurately recognize distracted driving.