基于混合Gamma分布的交叉口感应控制延误模型

为适应交叉口多态特征交通流,采用期望最大化算法,选取稠密性混合Gamma分布函数,准确拟合交叉口到达车头时距,统计交叉口进口到达交通量概率.以两相位全感应控制交叉口为例,根据交通量统计概率和信号相位状态获得每相位最大排队长度,并由此提出交叉口感应控制延误模型.在混合Gamma分布拟合精度不低于95%条件下的计算实例表明,基于混合Gamma分布的感应控制延误与实际结果更为吻合.

Intersection inductive control delay model based on mixture Gamma distribution

For adapting to the multistate characteristics of the intersection traffic flow,the intersection arrival headway is accurately fitted by using the expectation maximization algorithm and the dense mixture Gamma distribution function,and the probabilities of arrival traffic volume at intersection entrance are counted.Taking a two-phase full-actuated controlled intersection as the example,the maximum queue length of each phase can be obtained on the basis of statistics probabilities of traffic volume and states of signal phase.In this case,the delay model of actuated controlled intersection is proposed.Example calculation shows that,under the condition of the fitting accuracy of mixture Gamma distribution beyond 95%,the actuated control delay based on mixture Gamma distribution is more consistent with actual results.