基于遗传算法的交通信号机器学习控制方法

通过对到达车辆数目的模糊分类,将交通信号控制方案以不同规则集的形式实施,根据实际控制效果利用遗传算法对规则集进行改进,形成了一种具有机器学习能力的单路口交通信号新控制方法.经过仿真实验,对该方法的控制效果与定时控制和感应控制进行了比较,仿真实验的结果说明该方法的控制效果明显优于传统控制方式.

Machine-Learning Traffic Signal Control Approach Based on Genetic Algorithm

The traffic signal control schemes are put in the form of rule-sets into use through fuzzy classifying the arrived cars in this paper. Genetic algorithm is applied to improve the rule-sets according to the effect of actual controlling. These procedures form a new machining-learning traffic signal control approach for isolated intersection. After simulating, the control effect of this new approach with fixed-time control method and actuated control method are compared. The result of simulating illustrates that the effect of the new approach is obviously better than the traditional ones.