基于人工势能场的跟驰模型

Car-following model based on artificial potential field

为提升跟驰模型的模拟精度,更加有效地刻画跟驰行为特性,从车辆的实际运行规律出发,将行驶过程中的各种行为归纳为效率与安全2种因素的相互作用.借鉴人工势能场的基本思想,将这2种因素抽象为驾驶员受到的驱动力和阻碍力,进而建立相应的跟驰模型.根据普遍的驾驶行为特性,将车辆运行状态分为减速停车、起动加速和常态行驶3类,并据此对实测数据予以划分.利用分类后的实测数据,针对3类行驶状态分别对模型参数加以标定,并将标定后的模型与经典的GM模型进行比较.结果表明,相对于GM模型,模拟精度平均提升10%以上.

For improving the simulation accuracy of car-following model and describing the characteristics of car-following driving behavior, based on the driving experience, the actions in driving process are summarized into the interaction of efficiency and safety. Taking the basic idea of artificial potential fields for reference, these two factors are abstracted to be attraction and repulsion perceived by the driver, and then the car-following model is established. The three types of vehicles running is introduced, i.e. decelerating state, accelerating state and normal driving state. And the measured data are classified based on this. The classified measured data are used to calibrate the pa- rameters of model in different states, which are then compared with the classical GM （general motor） model. The result shows that the simulation accuracy is improved over ten percents.