驻足观望还是捷足先登？行人和无人驾驶车冲突演化博弈分析

doi:10.16381/j.cnki.issn1003-207x.2021.1576

摘要/Abstract

摘要：

无人驾驶的全面实现是一个循序渐进的过程，其间不可避免地会出现多形态的交通主体混行的情况。本文考虑了无信号路口处无人驾驶车辆与行人混行下的过街行为，引入了行人过路信任、无人驾驶车冲突损失溢出和行人过路信号交流等概念。从演化博弈的视角构建了双方冲突博弈矩阵，建立了双方交互演化的动力学模型，并分析得到了双方过街行为的演化稳定策略。结果表明：行人和无人驾驶车的策略选择与行人对无人驾驶技术的信任程度和无人驾驶车冲突损失溢出效应密切相关，在不同参数条件下，系统存在多种稳定均衡。此外，人车信号交流机制的引入可以在一定程度上消除由技术不信任带来的负面影响，促进博弈向目标方向演化。最后，本文通过数值模拟验证了模型的有效性，并给出了促进无人驾驶发展的管理启示。

关键词: 人车冲突, 演化博弈, 无人驾驶, 信任, 交流

Abstract:

Over recent years， traffic authorities in many countries have begun to approve autonomous vehicles for real-world road testing， and autonomous vehicles are gradually becoming one of the real travel options. However， the full realization of driverless is a gradual process， during which there will inevitably be a mixed situation of multiple forms of traffic. The most complex issue in this context is the interaction with pedestrians， a vulnerable traffic segment. Their movement trajectories are rather random and their actions are difficult to predict， which may lead to an increase in traffic accidents and cause considerable difficulties in traffic management. Considering that autonomous vehicles are different from traditional vehicles， it takes time for pedestrians to accept and adapt to autonomous vehicles， and autonomous vehicles also need to constantly update and optimize their strategies in the interaction with other traffic agents. Therefore， it is an urgent and important issue to analyze the pedestrian-vehicle interaction mechanism under driverless scenario from the perspective of evolutionary game and propose appropriate solutions to improve the traffic efficiency. Specifically， the crossing behavior of autonomous vehicles and pedestrians at unsignalized intersections is analyzed， and concepts are introduced such as pedestrians’ trust of driverless， loss spillover of autonomous vehicle conflict， and communication between pedestrians and vehicles. From the perspective of evolutionary game theory， a game matrix between the two parties is constructed. Then a dynamic model of the mutual evolution was established， and conclusively the stable strategies of the two parties' crossing behavior were obtained. The results show that the strategy choice of pedestrians and autonomous vehicles is closely related to the level of pedestrians’ trust of driverless and loss spillover of autonomous vehicle conflict. Under different parameter conditions， the system has multiple stable equilibriums. Besides， the introduction of the communication mechanism can eliminate the negative effects brought about by technological distrust to a certain extent， and promote the evolution of the game in the direction of the goal. Finally， the validity of the model is verified through simulation and gives management implications to promote the development of driverless. Managerial insights are provided from three perspectives： technology advocacy， media coverage and standard setting， with the aim of shedding new lights on the development of the autonomous vehicles industry.

Key words: conflicts between pedestrians and vehicles, evolutionary game theory, autonomous vehicles, trust, communication

中图分类号:

姚旭生,马寿峰,叶顺强,凌帅. 驻足观望还是捷足先登？行人和无人驾驶车冲突演化博弈分析[J]. 中国管理科学, 2024, 32(7): 138-149.

Xusheng Yao,Shoufeng Ma,Shunqiang Ye,Shuai Ling. Yield or Go?Evolutionary Game Analysis of Pedestrian and Autonomous Vehicle Conflicts[J]. Chinese Journal of Management Science, 2024, 32(7): 138-149.

图/表 10

表1

变量/参数定义"

符号	含义
$P, V$	行人/无人车过路基础收益
$R, S$	行人/无人车过路冲突损失
$M, N$	行人/无人车过路等待损失
$J, K$	行人/无人车过路互让损失
$T$	行人过路信任因子
$E$	无人驾驶车过路冲突损失溢出因子
$U$	行人过路信号交流收益

表1

表2

单个行人和单个无人驾驶车博弈的支付矩阵"

行人/无人驾驶车	通过 $(G)$	让步 $(Y)$
通过 $(G)$	$T P - R, V - S E$	$T P, V - N$
让步 $(Y)$	$P - M, V$	$P - J, V - K$

表2

表3

各均衡点对应雅克比矩阵的行列式的值和迹"

均衡点	$D e t (J)$	$T r (J)$
(0,0)	$(T P - P + J) K$	$T P - P + J + K$
(0,1)	$(R - M - T P + P) K$	$- R - M - T P + P - K$
(1,0)	$(T P - P + J) (S E - N)$	$- T P - P + J - (S E - N)$
(1,1)	$(R - M - T P + P) (S E - N)$	$R - M - T P + P + S E - N$
$(x d, y d)$	$- K S E - N T P - P + J R - M - T P + P S E - N + K R - M + J$	$0$

表3

图1

表4

考虑人车互动的单个行人和单个无人驾驶车博弈的支付矩阵"

行人/无人驾驶车	通过 $(G)$	让步 $(Y)$
通过 $(G)$	$T P - R, V - S E$	$T P + U, V - N$
让步 $(Y)$	$P - M + U, V$	$P - J, V - K$

表4

表5

演化动态系统（II）各均衡点对应雅克比矩阵的行列式的值和迹"

均衡点	$D e t (J)$	$T r (J)$
(0,0)	$(T P - P + J + U) K$	$T P - P + J + K + U$
(0,1)	$(R - M - T P + P + U) K$	$- R - M - T P + P + U - K$
(1,0)	$(T P - P + J + U) (S E - N)$	$- T P - P + J + U - (S E - N)$
(1,1)	$(R - M - T P + P + U) (S E - N)$	$R - M - T P + P + U + S E - N$
$(x d U, y d U)$	$- (K + U) (S E - N) (T P - P + J + U) (R - M - T P + P) (S E - N + K + U) (R - M + J + U)$	$0$

表5

表6

演化动态系统（II）均衡点类型判别表"

均衡点	$D e t (J)$	$T r (J)$	类型
(0,0)	+	+	不稳定点
(0,1)	+	－	稳定点（ESS）
(1,0)	+	－	稳定点（ESS）
(1,1)	+	+	不稳定点
$(x d U, y d U)$	－	0	鞍点

表6

图2

图3

图4

参考文献 27

1	胡明伟，张芷铭，陈湘生. 智能网联车辆混行交通流效益研究［J］. 系统仿真学报， 2021， 33（9）： 2270-2278.
	Hu M W， Zhang Z M， Chen X S. Research on benefits of mixed traffic flow of intelligent connected vehicles［J］. Journal of System Simulation， 2021， 33（9）： 2270-2278.
2	孙泽，贾斌，李新刚. 基于元胞自动机的行人和机动车相互干扰机理研究［J］. 物理学报， 2012， 61（10）： 76-83.
	Sun Z， Jia B， Li X G. The study of the interference between pedestrians and vehicles based on cellular automaton mode［J］. Acta Physica Sinica， 2012， 61（10）： 76-83.
3	Khallouk A， Echab H， Ez-Zahraouy H， et al. Traffic flow behavior at un-signalized intersection with crossings pedestrians［J］. Physics Letters A， 2018， 382（8）： 566-573.
4	韦艳芳，时伟，邝华，等. 无信号交叉口行人过街决策行为分析［J］. 上海大学学报（自然科学版）， 2013， 19（3）： 315-318.
	Wei Y F， Shi W， Kuang H， et al. Analysis of pedestrian crossing decision-making behaviors at unsignalized intersections［J］. Journal of Shanghai University （Natural Science Edition）， 2013， 19（3）： 315-318.
5	Xin X， Jia N， Ma S， et al. Empirical and simulation study of traffic delay at un-signalized crosswalks due to conflicts between pedestrians and vehicles［J］. Transportmetrica B：Transport Dynamics， 2019， 7（1）： 637-656.
6	Zhou R， Horrey W J. Predicting adolescent pedestrians' behavioral intentions to follow the masses in risky crossing situations［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2010， 13（3）： 153-163.
7	Onelcin P， Alver Y. Illegal crossing behavior of pedestrians at signalized intersections： Factors affecting the gap acceptance［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2015， 31： 124-132.
8	Althoff M， Stursberg O， Buss M. Model-based probabilistic collision detection in autonomous driving［J］. IEEE Transactions on Intelligent Transportation Systems， 2009， 10（2）： 299-310.
9	Luo Y， Cai P， Bera A， et al. Porca： Modeling and planning for autonomous driving among many pedestrians［J］. IEEE Robotics and Automation Letters， 2018， 3（4）： 3418-3425.
10	Cui L， Hu J， Park B B， et al. Development of a simulation platform for safety impact analysis considering vehicle dynamics， sensor errors， and communication latencies： Assessing cooperative adaptive cruise control under cyber attack［J］. Transportation Research Part C： Emerging Technologies， 2018， 97： 1-22.
11	刘晓波，鲁工圆，郑芳芳，等. 自动驾驶交通系统的协同管控技术［J］. 科学通报， 2020， 65（6）： 434-441.
	Liu X B， Lu G Y， Zheng F F， et al. Coordinated management and control of autonomous traffic systems［J］. Chinese Science Bulletin， 2020， 65（6）： 434-441.
12	Palmeiro A R， Kint S V D， Vissers L， et al. Interaction between pedestrians and automated vehicles： A wizard of oz experiment［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2018， 58： 1005-1020.
13	Rad S R， Correia G H， Hagenzieker M. Pedestrians’ road crossing behaviour in front of automated vehicles： Results from a pedestrian simulation experiment using agent-based modelling［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2020， 69： 101-119.
14	Zandi B， Singer T， Kobbert J， et al. International study on the importance of communication between automated vehicles and pedestrians［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2020， 74： 52-66.
15	魏丽英，崔裕枫，李东莹. 基于演化博弈论的行人与机动车冲突演化机理研究［J］. 物理学报， 2018， 67（19）： 37-49.
	Wei L Y， Cui Y F， Li D Y. Evolution mechanism of conflict between pedestrian and vehicle based on evolutionary game theory［J］. Acta Physica Sinica， 2018， 67（19）： 37-49.
16	林徐勋，袁鹏程，霍良安. 动态演化博弈视角下的“中国式过马路”问题研究［J］. 运筹与管理， 2016， 25（6）： 230-235.
	Lin X X， Yuan P C， Huo L A. “Chinese style road crossing” under the perspective of dynamic evolutionary game theory［J］. Operations Research and Management Science， 2016， 25（6）： 230-235.
17	Madigan R， Nordhoff S， Fox C， et al. Understanding interactions between automated road transport systems and other road users： A video analysis［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2019， 66： 196-213.
18	Camara F， Dickinson P， Fox C. Evaluating pedestrian interaction preferences with a game theoretic autonomous vehicle in virtual reality［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2021， 78： 410-423.
19	崔裕枫. 基于演化博弈论的行人与机动车冲突仿真模型研究［D］. 北京：北京交通大学学位论文， 2018.
	Cui Y F. Simulation study on conflict between pedestrian and vehicle based on evolutionary game theory［D］. Beijing： Thesis of Beijing Jiaotong University， 2018.
20	谷宁. 自动驾驶车辆交通需求预测模型研究［D］. 天津：河北工业大学学位论文， 2019.
	Gu N. Research on traffic demand forecasting model of autonomous vehicles［D］. Tianjin： Thesis of Hebei University of Technology， 2019.
21	Jayaraman S K， Creech C， Tilbury D M， et al. Pedestrian trust in automated vehicles： Role of traffic signal and av driving behavior［J］. Frontiers in Robotics and AI， 2019（6）： 117.
22	Friedman D. Evolutionary games in economics［J］. Econometrica： Journal of the Econometric Society， 1991，59（3）： 637-666.
23	Matthews M， Chowdhary G， Kieson E. Intent communication between autonomous vehicles and pedestrians［R］. Working Paper， arXiv， 2017.
24	Palmeiro A R， van der Kint S， Vissers L， et al. Interaction between pedestrians and automated vehicles： A Wizard of Oz experiment［J］. Transportation Research Part F： Traffic Psychology & Behaviour， 2018， 58： 1005-1020.
25	Lee Y M， Madigan R， Giles O， et al. Road users rarely use explicit communication when interacting in today’s traffic： Implications for automated vehicles［J］. Cognition， Technology & Work， 2021， 23： 367-380.
26	Chang C M， Toda K， Sakamoto D， et al. Eyes on a car： An interface design for communication between an autonomous car and a pedestrian［C］//Proceedings of the 9th International Conference on Automotive User Interfaces and Interactive Vehicular Applications， Oldenburg Germany， September 24-27 ，ACM Digital Library， 2017： 65-73.
27	Lagström T， Malmsten Lundgren V. AVIP-Autonomous vehicles' interaction with pedestrians： An investigation of pedestrian-driver communication and development of a vehicle external interface［D］. Gothenburg： Chalmers University of Technology， 2016.

[1]	杨志. 公众参与交互型跨界污染治理补偿的四方博弈[J]. 中国管理科学, 2024, 32(8): 261-273.
[2]	孙向彦,曲薪池. 药占比管制下过度医疗行为多阶段动态演化研究[J]. 中国管理科学, 2024, 32(8): 308-321.
[3]	刘钻扩,李宇,李守伟,邸凯旋,刘长玉,刘荣坤. 基于融资模式视角的第三方市场合作稳定策略研究[J]. 中国管理科学, 2024, 32(8): 36-49.
[4]	邵志国,李梦笛,韩传峰,孟令鹏,吴启迪. 基于演化博弈的建筑垃圾处理协同机制及仿真[J]. 中国管理科学, 2024, 32(7): 324-334.
[5]	梁雁茹,徐建中. 企业生态创新驱动系统激励政策优化研究[J]. 中国管理科学, 2024, 32(6): 281-289.
[6]	丁黎黎,张凯旋,王垒. 公众情绪如何影响工业共生链企业减排？[J]. 中国管理科学, 2024, 32(6): 312-322.
[7]	方国昌,何宇,田立新. 碳交易驱动下的政企碳减排演化博弈分析[J]. 中国管理科学, 2024, 32(5): 196-206.
[8]	丁黎黎,郭志蒙,白雨,王萍. 公平偏好对碳捕捉与封存项目产业化发展的影响分析[J]. 中国管理科学, 2024, 32(4): 187-197.
[9]	陈凤娇,刘德海. 重大传染病多渠道疫情信息发布模式的演化路径分析[J]. 中国管理科学, 2024, 32(4): 237-249.
[10]	赵丹,严啸宸,汪和平,李艳. 双积分政策下汽车企业合作创新演化博弈分析[J]. 中国管理科学, 2024, 32(4): 279-292.
[11]	景熠,曹柳,张闻秋. 考虑多元行为策略的地方政府大气治理四维演化博弈分析[J]. 中国管理科学, 2024, 32(4): 306-314.
[12]	李波,张春燕,张俊飚. 食品企业质量安全意识提升的演化博弈逻辑[J]. 中国管理科学, 2024, 32(4): 315-324.
[13]	杨松,张言彩,王爱峰. 多主体参与下食品安全社会共治演化博弈稳定性[J]. 中国管理科学, 2024, 32(4): 325-334.
[14]	张佩,白冰,赵紫威. 世界级竞争力集群内多方主体协同行为的演化博弈研究[J]. 中国管理科学, 2024, 32(4): 66-75.
[15]	孙睿,何大义,苏汇淋. 基于演化博弈的区块链技术在供应链金融中的应用研究[J]. 中国管理科学, 2024, 32(3): 125-134.