【摘要】 现有的干线自适应协调信号控制方法大都基于线圈或浮动车检测数据，控制效益受限于线圈布设密度、数据质量以及浮动车的渗透率等因素。近年，电警卡口设备被广泛应用于中国城市道路车辆违章和公共安全管理，可实时记录每一辆车通过交叉口停车线的时刻、所在车道及其身份信息，具有布设密度高、全样本检测、可还原车辆路径等优势，为干线自适应协调信号控制提供了新的数据支撑。基于此，提出一种基于电警数据的干线自适应协调信号控制优化方法：首先，通过充分挖掘电警数据蕴含的交通流信息，利用时空关联关系实时预测流向流量和路段行程时间，将其作为控制输入参数；其次，在经典的MULTIBAND模型基础上，提出适用于自适应控制的改进MULTIBAND模型，实现相位相序、公共周期、相位差等参数的自适应优化；最后，考虑到协调方案频繁切换可能带来的负面影响，设计了方案切换的判断逻辑和过渡算法以减少方案过渡的负面影响。仿真验证结果表明：流向流量和行程时间预测误差分别为12.66%和5.03%;在支路汇入流量较低和较高时，改进MULTIBAND模型相对于经典MULTIBAND模型的延误分别降低了5.2%和15.3%;在需求波动场景下，相较于基于改进MULTIBAND模型的定时协调控制和基于车头时距规则的感应协调控制，所提出的自适应控制方法可使干线整体的车均延误分别减少18.1%和15.1%,平均停车次数分别减少19.5%和15.4%。更多还原

【Abstract】 Many existing methods for arterial adaptive coordinated signal control are based on induction coil or probe vehicle detection data, and the control benefits are limited by factors, such as coil deployment density, data quality, and the penetration rate of probe vehicles. In recent years, E-police is widely used in China’s urban road-vehicle violation and public safety management. E-police can record the moment of each vehicle passing an intersection stop line, its lane, and its identity information in real time. With advantages of high deployment density, full sample detection, and reversible vehicle path, E-police provides new data support for arterial adaptive coordinated signal control. In this study, an optimization method for arterial adaptive coordinated signal control was developed based on the E-police data. First, by fully mining the traffic flow information contained in the E-police data, real-time prediction results of movement flow and link travel time using spatial-temporal correlation were used as the control input parameter. Second, an improved MULTIBAND model for adaptive control based on the classical MULTIBAND model was established to achieve the adaptive optimization of phase sequence, common cycle, offset and other parameters. Finally, considering the possible negative impact of frequent transitions of coordination schemes, a judgment logic of scheme switching and transition algorithm were designed to minimize the negative impact of scheme transition. The simulation validation results show that the prediction errors of movement flow and travel time are 12.66% and 5.03%, respectively. The delays of the improved MULTIBAND model relative to the classical MULTIBAND model decrease by 5.2% and 15.3% when the merging flows of branch roads are low and high, respectively. In the demand fluctuation scenario, compared with the fixed-timing coordinated control based on the improved MULTIBAND model and gap-based actuated coordinated control, the proposed adaptive control method can decrease the average vehicle delays of arterial by 18.1% and 15.1%, respectively, and the average number of stops by 19.5% and 15.4%, respectively.更多还原