A Time-of-the-day Partitioning Method for Traffic Signal Control Based on Key Intrinsic Mode Functions
-
摘要: 交通信号控制是缓解城市交通拥堵的重要手段,时段划分是信号灯控交叉口多时段控制的基础,合理的划分方法有助于提高信号控制效率。对于固定配时的信号灯控交叉口,传统时段划分方法主要借助于路口历史交通流量数据,依据人工经验或者简单聚类算法,直接进行时段划分,未能充分考虑交通流的时序性和随机性问题,不利于交通控制整体效益。综合考虑交通流中随机因素和时序性对时段划分的影响,本文研究了基于经验集合模态分解和有序聚类的时段划分方法。利用集合经验模态分解处理交叉口流量数据,提取了若干个本征模态函数及1个余项。借助皮尔逊相关系数分析原始流量数据、本征模态函数、余项这三者之间的关系,优选与原始流量相关性最高的本征模态函数或余项作为交通流的关键成分,使用关键成分代替流量数据进行有序聚类,完成时段划分。通过寻找不同分割个数下最小损失值突变点,获取最佳分割数,并得到最佳方案。以广东省中山市一个路口为案例对本文提出的时段划分方法进行算例分析,VISSIM仿真结果表明:①相比于现状,提出的方法在工作日和非工作日分别能提高路口通过车辆数11.32%和2.62%,缩短排队长度18.67%和12.02%;②非工作日车均延误减少6.80%,停车延误减少5.87%,工作日车均延误和停车延误变化不大。Abstract: Traffic signal control is an important tool to relieve urban traffic congestion and time-of-the-day partition is the basis for optimizing multi-period signal control at isolated signalized intersections in that a proper partition can significantly improve the efficiency of traffic control. For an intersection with a fixed-timing signal control strategy, traditional methods for time-of-the-day partition are usually based on experiences or simple clustering algorithms. These methods use historical traffic flow data to directly divide a day into several time periods, which fail to consider the stochasticity of traffic flow and the regularity of time sequence and lead to no contributions to the overall effectiveness of traffic control. To overcome this problem, this study proposes a new method for time-of-the-day partition, which uses an ensemble empirical mode decomposition (EEMD) and a fisher clustering algorithm. The intrinsic mode function (IMF) and corresponding residual from traffic flow data are extracted using EEMD. The Pearson correlation coefficient is calculated to analyze the relationship between the IMF, the residual, and the original traffic flow. The IMF or the residual that gives the highest correlation coefficient is identified as the key component, which replaces the traffic flows in the fisher clustering and partitioning process. The optimal number of clusters is determined by identifying the elbow point of the minimum loss values with different numbers of clusters, and the optimal time-of-the-day partition plan is obtained. A case study based on an intersection in the City of Zhongshan, Guangdong Province, is conducted to verify the proposed method. Simulations are carried out using the VISSIM software and study results show that ①Compares to the current situation, the proposed method can increase the number of vehicles going through the intersections by about 11.32% and 2.62% and can reduce the queue length by about 18.67% and 12.02% on weekdays and weekends, respectively. ②The proposed method also can reduce the average vehicle delay by 6.80% and the stopped delay by 5.87% at weekends, but cannot change both much during weekdays.
-
表 1 交叉口相位相序方案
Table 1. Phase sequence at the intersection
相序 相位 第1相位 南北直行 第2相位 南北左转 第3相位 东西直行 第4相位 东西左转 表 2 模型输入
Table 2. Inputs of models
算法名称 输入变量 Fisher有序聚类 路口总流量 EEMD-Fisher有序聚类 经EEMD分解后与路口总流量相关性最高的IMF分量 表 3 损失函数拐点汇总
Table 3. Turning point of loss function
日期 算法 拐点值 最终聚类数 Fisher 3.99 4 工作日 EEMD-Fisher-A 3.55 4 EEMD-Fisher-B 3.52 6 Fisher 2.71 8 非工作日 EEMD-Fisher-A 2.82 4 EEMD-Fisher-B 2.49 4 表 4 工作日现状时段划分及信号控制方案
Table 4. Current situation TOD and signal control plan on weekdays
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 39 15 26 20 100 07:00:00—08:59:59 78 24 39 27 168 09:00:00—11:59:59 64 22 32 26 144 12:00:00—13:59:59 47 20 33 20 120 14:00:00—16:59:59 64 22 32 26 144 17:00:00—18:59:59 78 24 39 27 168 19:00:00—22:00:00 47 20 33 20 120 表 5 工作日Fisher聚类时段划分及信号控制方案
Table 5. Fisher cluster TOD and signal control plan on weekdays
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 26 20 29 20 95 07:00:00—18:59:59 56 24 59 41 180 19:00:00—22:00:00 53 26 63 38 180 表 6 工作日EEMD-Fisher-A聚类时段划分及信号控制方案
Table 6. EEMD-Fisher-A TOD and signal control plan on weekdays
分割点 g1 g2 g3 g4 C 06:00:00—08:59:59 60 25 53 38 176 09:00:00—18:59:59 43 22 58 37 160 19:00:00—22:00:00 56 24 59 41 180 表 7 工作日EEMD-Fisher-B聚类时段划分及信号控制方案
Table 7. EEMD-Fisher-B TOD and signal control plan on weekdays
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 26 20 29 20 95 07:00:00—08:59:59 64 25 51 36 176 09:00:00—19:59:59 54 28 53 35 170 20:00:00—22:00:00 40 20 43 37 140 表 8 非工作日现状时段划分及信号控制方案
Table 8. Current situation TOD and signal control plan on weekends
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 27 16 38 19 100 07:00:00—16:59:59 73 27 34 26 160 17:00:00—18:59:59 60 27 41 32 160 19:00:00—22:00:00 47 20 33 20 120 表 9 非工作日Fisher聚类时段划分及信号控制方案
Table 9. Fisher cluster TOD and signal control plan on weekends
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 27 15 32 15 89 07:00:00—07:59:59 52 19 50 24 145 08:00:00—09:59:59 53 22 48 22 145 10:00:00—18:59:59 54 22 53 26 155 19:00:00—22:00:00 36 16 41 24 117 表 10 非工作日EEMD-Fisher-A聚类时段划分及信号控制方案
Table 10. EEMD-Fisher-A TOD and signal control plan on weekends
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 27 15 32 15 89 07:00:00—09:59:59 52 19 50 24 145 10:00:00—18:59:59 54 22 53 26 155 19:00:00—22:00:00 36 16 41 24 117 表 11 非工作日EEMD-Fisher-B聚类时段划分及信号控制方案
Table 11. EEMD-Fisher-B TOD and signal control plan on weekends
分割点 g1 g2 g3 g4 C 06:00:00—06:59:59 27 15 32 15 89 07:00:00—09:59:59 53 20 48 24 145 10:00:00—19:59:59 50 21 50 34 155 20:00:00—22:00:00 32 15 39 31 117 表 12 工作日平均排队长度和通过车辆数
Table 12. Queue length and vehicles passing the intersections on weekdays
指标 平均排队长度/m 通过车辆数/veh 现状 22.49 61 425.20 Fisher 20.43 63 648.80 (-9.16%) (+3.62%) EEMD-Fisher-A 17.99 66 002.80 (-20.01%) (+7.45%) EEMD-Fisher-B 18.29 68 379.40 (-18.67%) (+11.32%) 表 13 工作日车均延误和停车延误
Table 13. Average vehicle delay and stopped delay on weekdays
指标 车均延误/s 停车延误/s 现状 73.77 59.65 Fisher 80.49 66.44 (+9.11%) (+11.38%) EEMD-Fisher-A 78.70 65.10 (+6.68%) (+9.14%) EEMD-Fisher-B 73.53 60.07 (-0.33%) (+0.70%) 表 14 非工作日平均排队长度和通过车辆数
Table 14. Queue length and vehicles passing the intersections on weekends
指标 平均排队长度/m 通过车辆数/veh 现状 14.64 63 477.20 Fisher 13.65 63 801.00 (-6.76%) (+0.51%) EEMD-Fisher-A 12.70 64 090.40 (-13.25%) (+0.97%) EEMD-Fisher-B 12.88 65 138.60 (-12.02%) (+2.62%) 表 15 非工作日车均延误和停车延误
Table 15. Average vehicle delay and stopped delay on weekends
指标 车均延误/s 停车延误/s 现状 62.02 49.57 Fisher 61.40 48.90 (-1.00%) (-1.35%) EEMD-Fisher-A 60.20 48.01 (-2.93%) (-3.15%) EEMD-Fisher-B 57.80 46.66 (-6.80%) (-5.87%) -
[1] 罗舒琳, 张存保, 张泰文, 等. 面向常发拥堵点的主动交通诱导方法面向常发性拥堵的城市局部路网韧性评价与分析[J]. 交通信息与安全, 2021, 39(5): 68-75. doi: 10.3963/j.jssn.1674-4861.2021.05.009LUO S L, ZHANG C B, ZHANG T W, et al. Active traffic guidance method for recurrent congestion points[J]. Journal of Transport Information and Safety, 2021, 39(5): 68-75. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2021.05.009 [2] CHEN P, ZHENG N, SUN W L, et al. Fine-tuning time-of-day partitions for signal timing plan development: Revisiting clustering approaches[J]. Transportmetrica A: Transport Science, 2019, 15(2): 1195-1213. doi: 10.1080/23249935.2019.1571536 [3] 别一鸣, 姜凯, 汤茹茹, 等. 考虑方案过渡影响的单点交通控制时段划分方法[J]. 吉林大学学报(工学版), 2019, 49(6): 1844-1851. doi: 10.13229/j.cnki.jdxbgxb20181119BIE Y M, JIANG K, TANG R R, et al. Time of interval partition for traffic control at isolated intersection considering impacts of plan transition[J]. Journal of Jilin University(Engineering and Technology Edition), 2019, 49(6): 1844-1851. (in Chinese) doi: 10.13229/j.cnki.jdxbgxb20181119 [4] WANG G Z, QIN W, WANG Y H. Cyclic weighted k-means method with application to time-of-day interval partition[J]. Sustainability, 2021, 13(9): 4796-4809. doi: 10.3390/su13094796 [5] RATROUT N. Subtractive clustering-based k-means technique for determining optimum time-of-day breakpoints[J]. Journal of Computing in Civil Engineering, 2011, 25(5): 380-387. doi: 10.1061/(ASCE)CP.1943-5487.0000099 [6] SONG X, LI W J, MA D F, et al. An enhanced clustering-based method for determining time-of-day breakpoints through process optimization[J]. IEEE Access, 2018(6): 29241-29253. [7] 熊睿成, 袁淑芬, 安成川, 等. 单点交叉口信号控制方案时段划分方法[C]. 第十四届中国智能交通年会, 北京: 中国智能交通协会, 2019.XIONG R C, YUAN S F, AN C C, et al. Time-of-day signal plan division methods at isolated intersections[C]. The 14th Annual Conference on Intelligent Transportation in China, Beijing: China Intelligent Transportation Systems Association, 2019. (in Chinese) [8] MA D F, LI W J, SONG X, et al. Time-of-day breakpoints optimization through recursive time series partitioning[J]. IET Intelligent Transport Systems, 2019, 13(4): 683-692. doi: 10.1049/iet-its.2018.5162 [9] 李文婧, 孙锋, 李茜瑶, 等. 采用递归有序聚类的信号控制时段划分方法[J]. 浙江大学学报(工学版), 2018, 52(6): 1150-1156. https://www.cnki.com.cn/Article/CJFDTOTAL-ZDZC201806014.htmLI W J, SUN F, LI X Y, et al. Time-of-day breakpoints for traffic signal control using dynamic recurrence order clustering[J]. Journal of Zhejiang University(Engineering Science), 2018, 52(6): 1150-1156. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-ZDZC201806014.htm [10] GARCÍA-RÓDENAS R, LÓPEZ-GARCÍA M L, SÁNCHEZ-RICO M T, et al. A bilevel approach to enhance prefixed traffic signal optimization[J]. Engineering Applications of Artificial Intelligence, 2019, 84: 51-65. doi: 10.1016/j.engappai.2019.05.017 [11] 于德新, 田秀娟, 杨兆升. 基于改进FCM聚类的交通控制时段划分[J]. 华南理工大学学报(自然科学版), 2016, 44(12): 53-60. https://www.cnki.com.cn/Article/CJFDTOTAL-HNLG201612008.htmYU D X, TIAN X J, YANG Z S. Division of traffic control periods based on improved FCM clustering[J]. Journal of South China University of Technology(Natural Science Edition), 2016, 44(12): 53-60. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-HNLG201612008.htm [12] SHEN H, YAN J, LIU D G, et al. A new method for determination of time-of-day breakpoints based on clustering and image segmentation[J]. Canadian Journal of Civil Engineering, 2020, 47(8): 974-981. doi: 10.1139/cjce-2019-0153 [13] WAN L J, YU C H, WANG L, et al. Identification of time-of-day breakpoints based on trajectory data of probe vehicles[J]. Transportation Research Record, 2019, 2673(5): 538-547. [14] DU Z L, YAN X T, ZHU J Q, et al. Signal timing parameters estimation for intersections using floating car data[J]. Transportation Research Record, 2019, 2673(6): 189-201. doi: 10.1177/0361198119844756 [15] HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis[J]. Proceedings of the Royal Society of London. Series A: Mathematical, Physical Engineering Sciences, 1998, 454(1971): 903-995. [16] WU Z H, HUANG N E. Ensemble empirical mode decomposition: A noise-assisted data analysis method[J]. Advances in Adaptive Data Analysis, 2009, 1(1): 1-41. [17] ZHENG L J, YANG J, CHEN L, et al. Dynamic spatial-temporal feature optimization with ERI big data for Short-term traffic flow prediction[J]. Neurocomputing, 2020, 412: 339-350. [18] FENG B, XU J M, ZHANG Y G, et al. Multi-Step traffic speed prediction based on ensemble learning on an urban road network[J]. Applied Sciences, 2021, 11 (10): 4423-4438. [19] 周燕青. 基于流量动态的单交叉口多时段信号控制优化研究[D]. 重庆: 重庆交通大学, 2017.ZHOU Y Q. Research on multi-period signal optimization of signal intersection based on flow dynamic[D]. Chongqing: Chongqing Jiaotong University, 2017. [20] 周丽. 基于动态时段划分的交叉口信号控制模型与算法研究[D]. 济南: 山东大学, 2011.ZHOU L. Study on signal control model and algorithm at isolated intersection based on dynamic time-of-day[D]. Jinan: Shandong University, 2011. [21] 蒋金勇, 云美萍, 杨佩昆. 基于HCM2000延误模型的最佳周期时长估算公式[J]. 同济大学学报(自然科学版), 2009, 37(8): 1024-1028. https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ200908006.htmJIANG J Y, YUN M P, YANG P K. Optimal cycle length estimation equations based on delay models of HCM2000[J]. Journal of Tongji University(Natural Science), 2009, 37(8): 1024-1028. (in Chinese). https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ200908006.htm [22] 吴兵, 李晔. 交通管理与控制[M]. 北京: 人民交通出版社, 2020.WU B, LI Y. Traffic Management and Control[M]. Beijing: China Communications Press, 2020. (in Chinese).