一种基于分类回归树的无人车汇流决策方法

苏锑; 杨明; 王春香; 唐卫; 王冰

doi:10.16383/j.aas.2018.c160457

一种基于分类回归树的无人车汇流决策方法

doi: 10.16383/j.aas.2018.c160457

苏锑^1,,
杨明^2,3, ,,
王春香^1,,
唐卫^2,3,,
王冰^2,3,

1.
上海交通大学机器人所上海 200240
2.
上海交通大学自动化系上海 200240
3.
上海市北斗导航与位置服务重点实验室上海 200240

基金项目:

国家磁约束核聚变能研究专项 2012GB102002

国家自然科学基金 91420101

详细信息

作者简介:
苏锑上海交通大学智能车实验室硕士研究生.2014年获得河海大学学士学位.主要研究方向为协作驾驶算法和无人车队.E-mail:teenie_su@139.com

王春香上海交通大学机械与动力工程学院副教授.1999年在哈尔滨工业大学获得博士学位.主要研究方向为移动机器人, 自动驾驶及高级驾驶辅助.E-mail:wangcx@sjtu.edu.cn

唐卫上海交通大学自动化系博士研究生.2008年在南京理工大学获得学士学位, 2011年在上海航天技术研究院获得硕士学位.主要研究方向为协作驾驶和多智能体.E-mail:tangwei_327@163.com

王冰上海交通大学自动化系高级工程师.1997年在上海交通大学获得工学博士学位.主要研究方向为嵌入式控制, 工业自动化, 自动导引车.E-mail:bingwang@sjtu.edu.cn

通讯作者:
杨明上海交通大学智能车研究所主任, 自动化系教授.1999年和2003年在清华大学获得计算机科学硕士和博士学位.主要研究方向为自动驾驶, 协作驾驶, 移动机器人, 机器视觉和高精细地图.本文通信作者.E-mail:mingyang@sjtu.edu.cn

计量
- 文章访问数: 2969
- HTML全文浏览量: 527
- PDF下载量: 958
- 被引次数: 0
出版历程
- 收稿日期: 2016-06-12
- 录用日期: 2016-11-23
- 刊出日期: 2018-01-01

Classification and Regression Tree Based Traffic Merging for Method Self-driving Vehicles

SU Ti^1
,,
YANG Ming^{2,3
, ,},
WANG Chun-Xiang^1
,,
TANG Wei^{2,3
,},
WANG Bing^{2,3
,}

1.
Research Institute of Robotics, Shanghai Jiao Tong University, Shanghai 200240
2.
Department of Automation, Shanghai Jiao Tong University, Shanghai 200240
3.
Shanghai Key Lab of Navigation and Location Services, Shanghai 200240

Funds:

National Magnetic Conflnement Fusion Energy Research Project 2012GB102002

National Natural Science Foundation of China 91420101

More Information

Author Bio:
Master student at Shanghai Jiao Tong University. She received her bachelor degree from HoHai University in 2014. Her research interest covers cooperative driving and autonomous platooning

Associate professor at the School of Mechanical Engineering, Shanghai Jiao Tong University. She received her Ph. D. degree in mechanical engineering from Harbin Institute of Technology in 1999. Her research interest covers mobile robots, autonomous driving, and assistant driving

Ph. D. candidate in the Department of Automation, Shanghai Jiao Tong University. He received his bachelor degree in automation from Nanjing University of Science and Technology in 2008 and master degree in navigation guidance and control from Shanghai Academy of Spaceflight Technology in 2011. His research interest covers cooperative driving and multi-agents system

Senior engineer in the Department of Automation, Shanghai Jiao Tong University. He received his Ph. D. degree in mechanical engineering from Shanghai Jiao Tong University in 1997. His research interest covers embedded control, industrial automation, and automated guided vehicles

Corresponding author: YANG Ming Director of the Research Institute of Intelligent Vehicle Technology, professor in the Department of Automation, Shanghai Jiao Tong University. He received his master and Ph. D. degrees in computer sciences from Tsinghua University in 1999 and 2003, respectively. His research interest covers autonomous driving, cooperative driving, mobile robots, machine vision, and high deflnition map. Corresponding author of this paper

摘要

摘要: 决策规划是无人驾驶技术中的重要环节.由于道路结构变化或障碍物引起的车辆被动换道多采用基于逻辑规则或优化算法的决策方式.本文以通行量为优化目标，提出一种基于分类回归树（Classification and regression tree，CART）的汇流决策方法.依据交通流参数，选择大量具有代表性的车辆汇流场景.对场景中车辆的汇流决策序列进行编码，采用遗传算法搜索使得通行量最大的决策方案.将寻优获得的大量汇流决策序列作为样本，训练分类回归树.选取车辆自身信息及与周围车辆的关系等以描述环境特征，运用分类回归树描述环境特征与决策结果的映射关系，获得一种通行量最优的汇流决策方法.在软件中进行仿真实验，对比既有方法，基于分类回归树的汇流方法能够有效减少汇流行为对车流的扰动，在大流量情形下依旧能保持较高的通行效率.此外，该方法对实际实施中可能存在的环境感知误差，如定位误差，有一定的鲁棒性.
- 汇流决策 /
- 遗传算法 /
- 分类回归树 /
- 交通流仿真
Abstract: Decision-making and planning are important technologies of unmanned vehicle. Logical rule and optimization algorithm are commonly applied to passive merging strategy for road structure change or obstacles. A traffic merging strategy aiming to improve throughput is proposed in this paper. According to different traffic parameters, a large number of typical traffic merging scenarios are selected. For vehicles in different scenarios, decision sequences are encoded and optimal merging decision is obtained by genetic algorithm based on remainder stochastic sampling with replacement (RSSR). Those optimal decisions are used to train classification and regression tree (CART). Specifically, the environmental feature is described by vehicle state and relationship between other vehicles around. Then the relationship between environmental features and decision is modeled by classification and regression tree. Compared with the previous merging strategy it is shown by simulation that the merging strategy based on CART can effectively mitigate disturbance on traffic flow, brought by merging maneuver, and maintain a high through efficiency even in large flow circumstances. Moreover, this method is also rather robust to environmental perception errors, such as positioning error which may exist in implementation.
- Merging strategy /
- genetic algorithm /
- classification and regression tree (CART) /
- simulation of urban mobility (SUMO)
注释:

1) 本文责任编委张毅

HTML全文

图 1 汇流场景

Fig. 1 Merging scenario

下载: 全尺寸图片幻灯片

图 2 汇流决策时间序

Fig. 2 Merging decision sequence

下载: 全尺寸图片幻灯片

图 3 环境特征描述示例

Fig. 3 Environment feature description

下载: 全尺寸图片幻灯片

图 4 分类决策树结构

Fig. 4 Classification and regression tree

下载: 全尺寸图片幻灯片

图 5 SUMO仿真环境

Fig. 5 SUMO simulation environment

下载: 全尺寸图片幻灯片

图 6 平均流量1 400 veh/h下游平均流量

Fig. 6 1 400 veh/h, mean flow

下载: 全尺寸图片幻灯片

图 7 平均流量2 600 veh/h下游平均流量

Fig. 7 600 veh/h, mean flow

下载: 全尺寸图片幻灯片

图 8 汇流点上游平均速度

Fig. 8 Mean velocity of upstream

下载: 全尺寸图片幻灯片

图 9 平均流量1 400 veh/h定位误差对汇流效果影响

Fig. 9 Influence of positioning error on merging efficiency, mean flow 1 400 veh/h

下载: 全尺寸图片幻灯片

图 10 平均流量2 600 veh/h定位误差对汇流效果影响

Fig. 10 Influence of positioning error on merging efficiency, mean flow 2 600 veh/h

下载: 全尺寸图片幻灯片

表 1 环境特征变量描述

Table 1 Environment feature description

变量(Var)	描述
${D(i)}$	汇流车辆距离瓶颈点的距离
${V(i)}$	汇流车辆车速
${X(i-1)}$	汇流车辆与本车道中前车的纵向距离
${TTC(i-1)}$	汇流车辆与本车道中前车的预估碰撞时间
${X(k-1)}$	汇流车辆与汇入车道中前车的纵向距离
${TTC(k-1)}$	汇流车辆与汇入车道中前车的预估碰撞时间
${X(k)}$	汇入车道中后车与汇流车辆的纵向距离
${TTC(k)}$	汇入车道中后车与汇流车辆的预估碰撞时间

下载: 导出CSV

表 2 平均等待时间比较

Table 2 Mean waiting time comparison

汇流方案	到达过程	平均流量	最长等待时间	平均等待时间
汇流方案	到达过程	(veh/h)	(s)	(s)
CART	Poisson	1 200	0	0
CART	Poisson	1 800	0	0
CART	Poisson	2 400	1.1	0.02
CART	Constant	1 800	0	0
CART	Constant	2 400	1.4	0.021
PV	Poisson	1 200	88.3	20.5
PV	Poisson	1 800	203.2	119.9
PV	Poisson	2 400	554.8	345.1
PV	Constant	1 800	176.6	60.7
PV	Constant	2 400	198.4	101.3
OMS	Poisson	1 200	0	0
OMS	Poisson	1 800	0	0
OMS	Poisson	2 400	0.63	0.01
OMS	Constant	1 800	0	0
OMS	Constant	2 400	0.84	0.01

下载: 导出CSV

参考文献(17)

[1]	Kerner B S. Experimental features of self-organization in traffic flow. Physical Review Letters, 1998, 81(17):3797-3800 doi: 10.1103/PhysRevLett.81.3797
[2]	Krauß S. Microscopic Modeling of Traffic Flow:Investigation of Collision Free Vehicle Dynamics[Ph.D. dissertation], University of Cologne, Germany, 1998
[3]	Cao W J, Muka M, Kawabe T, Nishira H, Fujiki N. Merging trajectory generation for vehicle on a motor way using receding horizon control framework consideration of its applications. In:Proceedings of the 2014 IEEE Conference on Control Applications (CCA). Juan Les Antibes, France:IEEE, 2014. 2127-2134
[4]	Marinescu D, Čurn J, Bouroche M, Cahill V. On-ramp traffic merging using cooperative intelligent vehicles:a slot-based approach. In:Proceedings of the 15th International IEEE Conference on Intelligent Transportation Systems (ITSC). Anchorage, AK, USA:IEEE, 2012. 900-906
[5]	Rios-Torres J, Malikopoulos A, Pisu P. Online optimal control of connected vehicles for efficient traffic flow at merging roads. In:Proceedings of the 18th International Conference on Intelligent Transportation Systems (ITSC). Las Palmas, Spain:IEEE, 2015. 2432-2437
[6]	Awal T, Kulik L, Ramamohanrao K. Optimal traffic merging strategy for communication-and sensor-enabled vehicles. In:Proceedings of the 16th International IEEE Conference on Published of the Intelligent Transportation Systems-(ITSC). The Hague, Netherlands:IEEE, 2013.
[7]	Wang Z Y, Kulik L, Ramamohanarao K. Proactive traffic merging strategies for sensor-enabled cars. In:Proceedings of the 4th ACM International Workshop on Vehicular Ad Hoc Networks. New York, NY, USA:ACM, 2007. 39-48
[8]	陈思曼, 孟宪实, 马钧.匝道口智能车合流避撞模型及仿真研究.农业装备与车辆工程, 2016, 54(2):44-50 http://mall.cnki.net/magazine/Article/SDLG201602026.htm Chen Si-Man, Meng Xian-Shi, Ma Jun. Study on the model and simulation for confluence avoidance at ramp intersection. Agricultural Equipment and Vehicle Engineering, 2016, 54(2):44-50 http://mall.cnki.net/magazine/Article/SDLG201602026.htm
[9]	Kita H. A merging-giveway interaction model of cars in a merging section:a game theoretic analysis. Transportation Research Part A:Policy and Practice, 1999, 33(3-4):305-312 doi: 10.1016/S0965-8564(98)00039-1
[10]	Li L, Wen D, Yao D Y. A survey of traffic control with vehicular communications. IEEE Transactions on Intelligent Transportation Systems, 2014, 15(1):425-432 doi: 10.1109/TITS.2013.2277737
[11]	Li L, Wang F Y, Zhang Y. Cooperative driving at lane closures. In:Proceedings of the 2007 IEEE Intelligent Vehicles Symposium. Istanbul, Turkey:IEEE, 2007. 1156-1161
[12]	Li L, Wang F Y. Cooperative driving at blind crossings using intervehicle communication. IEEE Transactions on Vehicular Technology, 2006, 55(6):1712-1724 doi: 10.1109/TVT.2006.878730
[13]	Weng J X, Xue S, Yan X D. Modeling vehicle merging behavior in work zone merging areas during the merging implementation period. IEEE Transactions on Intelligent Transportation Systems, 2016, 17(4):917-925 doi: 10.1109/TITS.2015.2477335
[14]	Brindle A. Genetic Algorithms for Function Optimization[Ph.D. dissertation], University of Alberta, Canada, 1981.
[15]	Pei Y L, Dai L L. Study on intelligent lane merge control system for freeway work zones. In:Proceedings of the 2007 Intelligent Transportation Systems Conference. Seattle, WA, USA:IEEE, 2007. 586-591
[16]	Wegener A, Piórkowski M, Raya M, Hellbrück H, Fischer S, Hubaux J P. TraCI:an interface for coupling road traffic and network simulators. In:Proceedings of the 11th Communications and Networking Simulation Symposium. New York, NY, USA:ACM, 2008. 155-163
[17]	Batista G E A P A, Prati R C, Monard M C. A study of the behavior of several methods for balancing machine learning training data. ACM Sigkdd Explorations Newsletter, 2004, 6(1):20-29 doi: 10.1145/1007730