专用道公交优先干线协调分区判据方法研究

发布时间：2018-04-15 11:34

本文选题：专用道公交优先 + 驻站时间预测　；参考：《北方工业大学》2017年硕士论文

【摘要】：随着经济的日益发展,城市公共交通服务建设成为城市道路交通发展的重中之重。公交信号优先干线协调因其实施方便、效果突出等特点能很好的提高城市公共交通服务水平。在公交专用道连续性得到提升的前提下,进一步研究专用道公交干线优先控制方法将具有更强的实际应用价值。本文从专用道公交车辆驻站时间入手,结合专用道公交干线协调控制分区判据的研究,完善专用道公交社会车辆干线协调策略,具体主要分为以下三个部分:第一、专用道公交驻站时间预测。基于检测器数据经过数据预处理得到特定班次前后车到站时间间隔以及后车驻站时间的数据矩阵,采用多项式回归、指数函数回归、指数函数回归等措施对其回归拟合,采用皮尔逊系数对其回归精度筛选以确定预测模型,通过平均预测精度来衡量驻站时间预测精度,最后通过蒙特卡洛算法和均值聚类算法计算动态优化误差阈值,实现预测结果的动态优化。第二、专用道干线协调分区判据。首先,基于第一部分的驻站时间预测,结合现有的专用道公交车辆的进站平均速度和离站平均速度,以路段车辆检测器的检测时刻为准推算公交车到达下游交叉口的时间;其次,通过结合下游交叉口的现有配时策略,统计半小时内公交车的一次性通过率,通过对一次性通过率矩阵进行伪F统计量分析得到最佳分类数,将一维欧氏距离采用均值聚类算法分成最佳类别;最后,通过周期判断原则分析上下游交叉口间的分区关系,最终得到专用道公交的干线协调分区。第三、多时段干线协调下专用道公交优先。首先,基于谱聚类算法实现公交车辆的浮动加权处理得到体现公交优先的加权流量,并据此流量通过采用时间序列划分算法研究单点的多时段划分需求;实现多时段划分和周期计算中体现公交优先的目标;然后,依据干线协调优化算法,对干线交叉口信号控制进行优化,以达到最大协调带宽;最后,经过仿真验证该策略的有效性和可靠性。
[Abstract]:With the development of economy, the construction of urban public transport service has become the most important part of urban road traffic development.The coordination of bus signal priority trunk lines can improve the service level of urban public transportation because of its convenient implementation and outstanding effect.On the premise of improving the continuity of bus lanes, it will be of more practical value to further study the priority control method of bus trunk lines.In this paper, starting with the stop time of public transport vehicles in special lanes, combined with the study of the criterion of coordinated control zoning of bus trunk lines of exclusive lanes, the coordination strategy of social trunk lines of public transport in special lanes is improved. The main contents are as follows: first,The bus stop time forecast.Based on the data preprocessing of the detector data, the data matrix of the time interval between the arrival and arrival of the train before and after a specific shift and the time of the rear train stop are obtained. The polynomial regression, the exponential function regression, the exponential function regression and so on are used to fit the data matrix.The regression accuracy of Pearson coefficient is screened to determine the prediction model, the prediction accuracy is measured by the average prediction precision, and the dynamic optimization error threshold is calculated by Monte Carlo algorithm and mean clustering algorithm.Dynamic optimization of prediction results is realized.Second, the special track trunk line coordinates the partition criterion.First of all, based on the first part of the station time prediction, combined with the existing bus stop average speed and exit average speed, based on the detection time of the vehicle detector in the section of the road to calculate the time of the bus to the downstream intersection;Secondly, by combining the existing timing strategy of the downstream intersection, the one-off pass rate of the bus within half an hour is counted, and the best classification number is obtained by the pseudo-F statistic analysis of the one-off pass rate matrix.The one-dimensional Euclidean distance is divided into the best category by means of the mean value clustering algorithm. Finally, the partition relationship between the upstream and downstream intersections is analyzed according to the principle of periodic judgment, and the coordinated trunk line partition of the dedicated public transportation is finally obtained.Third, under the coordination of multi-time trunk lines, public transport is preferred.Firstly, based on the spectral clustering algorithm, the floating weight processing of public transport vehicles is realized to get the weighted flow which reflects the priority of public transport. According to this, the demand of multi-time division of a single point is studied by using the time series partition algorithm.Realization of multi-time division and cycle calculation to reflect the goal of bus priority; then, according to the trunk line coordination optimization algorithm to optimize the trunk intersection signal control, in order to achieve the maximum coordinated bandwidth; finally,The effectiveness and reliability of the strategy are verified by simulation.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.54

【参考文献】