论文部分内容阅读
为了提高交叉口运行效率,提出了一种基于胞映射的交叉口自学习模糊控制策略。该方法首先以交叉口各方向排队长度为状态构成状态空间,并进一步将该空间划分为离散的模糊胞元;然后通过研究交叉口交通状态在状态空间内各胞元间的跳转关系,分析交叉口系统的动态特性;最后以此为基础制定模糊控制规则,设计模糊控制器以确定交叉口的绿信比,并引入自学习策略对控制效果进行评估和持续改进控制器性能。基于北京市地安门的实测交通数据的仿真结果表明:该方法与固定配时控制与感应控制相比较,能够明显地减少交叉口各方向的排队长度。
In order to improve the operation efficiency of intersection, a cell-based intersection self-learning fuzzy control strategy is proposed. The method firstly constructs the state space with the queuing length in each direction of the intersection as a state, and further divides the space into discrete fuzzy cells. Then, by studying the jump relations among the cells in the state space of the traffic state at the intersection, Finally, based on this, the fuzzy control rules are formulated. The fuzzy controller is designed to determine the green signal-to-signal ratio at the intersection, and a self-learning strategy is introduced to evaluate the control effect and improve the controller performance continuously. The simulation results of the measured traffic data based on Beijing Dianemen show that compared with the fixed timing control and the inductive control, this method can obviously reduce the queuing length in all directions of the intersection.