论文部分内容阅读
以后车车速、前后车速差和车间距作为观察变量输入,驾驶人的驾驶状态作为隐含变量输出,利用隐马尔科夫模型提出了一种机动车驾驶人状态预测方法。首先筛选预测所需的观察状态序列,接着利用隐Markov求解预测时刻所有选中的观察状态序列出现的概率以及各观察状态序列和指定驾驶状态(隐含变量)同时出现的概率,最后利用条件概率将上述两者转化为驾驶人状态概率。为检验方法的预警性能,除考核预报的正确性外,定义了“预报度”衡量驾驶人不良状态概率为P时提前预警的时间。仿真结果表明,预测的驾驶状态变化趋势与PERCLOS监测结果一致,且能实现提前预报,P值越小,预报度越大。
After the vehicle speed, the speed difference between the front and the rear and the distance between the vehicles as the input of observation variables, the driver’s driving status is output as an implicit variable. A prediction method of vehicle driver’s status is proposed by using hidden Markov model. Firstly, the sequence of observation states needed for prediction is first filtered, and then the probability of occurrence of all selected observation states and the probability of simultaneous appearance of each observed state sequence and the specified driving state (hidden variables) are predicted by using hidden Markov. Finally, conditional probability The above two are transformed into the driver’s state probability. In order to test the method of early warning performance, in addition to the accuracy of the test forecast, the definition of “forecast ” to measure the probability of poor driver status P early warning time. The simulation results show that the predicted driving status change trend is consistent with the PERCLOS monitoring results and can predict ahead of time. The smaller the P value, the larger the forecasting degree.