作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2022, Vol. 48 ›› Issue (4): 299-306. doi: 10.19678/j.issn.1000-3428.0060237

• 开发研究与工程应用 • 上一篇    下一篇

V2V环境下具有组策略防护特性的虚拟交通灯

宋玮, 赵会奋, 蔡文钦, 周万强   

  1. 广东工业大学 计算机学院, 广州 510006
  • 收稿日期:2021-01-15 修回日期:2021-02-24 发布日期:2021-04-15
  • 作者简介:宋玮(1978—),女,讲师、博士,主研方向为博弈论算法、机制设计智能交通;赵会奋、蔡文钦、周万强,本科生。
  • 基金资助:
    广东省信息物理融合系统重点实验室项目(2016B030301008);广东工业大学教师国(境)外研修支持计划;省级大学生创新创业训练项目(S202011845159,S202011845160);校级大学生创新创业训练项目(xj202011845384)。

Group-Strategy-Proof Virtual Traffic Light under V2V Environment

SONG Wei, ZHAO Huifen, CAI Wenqin, ZHOU Wanqiang   

  1. Guangdong University of Technology, School of computer, Guangzhou 510006, China
  • Received:2021-01-15 Revised:2021-02-24 Published:2021-04-15

摘要: 在V2V环境下的虚拟交通灯可以通过车辆间直接交换的信息协商路权分配,且在设备获取相关信息时,车辆能够有策略地提供信息以获得优先路权。为适用于非可测量因素影响路权的场景,提出一种具有组策略防护特性的虚拟交通灯。通过将车辆提供的真实信息抽象为成本分摊与合作博弈,并设计组策略防护拍卖机制,利用Shapley值计算出每辆车的成本分摊作为车辆的支付。在此基础上,根据拍卖结果中真实的评价值建立绿灯信号,通过信号合并算法整合多次拍卖产生的绿灯信号,由此产生合理的路权分配。实验结果表明,该虚拟交通灯具有组策略防护特性,能够避免车辆形成虚报信息的联盟来获取利益,也能避免车辆通过虚报私有信息来获得路权优先权,与具有固定绿灯通行数量阈值的虚拟交通灯相比,组策略防护的虚拟交通灯在整体平均行驶时间以及高评价值车辆的平均行驶时间上均有一定改善。

关键词: V2V环境, 虚拟交通灯, 成本分摊合作博弈, 组策略防护拍卖机制, Shapley值, 平均行驶时间

Abstract: The Virtual Traffic Light (VTL) in a Vehicle-to-Vehicle (V2V) environment can negotiate the right-of-way allocation through the information directly exchanged between vehicles.When the equipment obtains relevant information, the vehicle can strategically provide information to obtain the priority right of way.To apply to a scene where unmeasurable factors affect the right of way, a virtual traffic light with group strategy protection characteristics is proposed.By abstracting the real information provided by vehicles into a cost allocation and cooperative game, a group strategy protection auction mechanism is designed, and the Shapley value is used to calculate the cost allocation of each vehicle as the payment of vehicles.On this basis, the green light signal is established according to the real evaluation value in the auction results, and the green light signal generated by multiple auctions is integrated through the signal merging algorithm to produce a reasonable right-of-way allocation.The experimental results show that the virtual traffic light has the characteristics of group strategy protection, which can prevent vehicles from forming an alliance of false information to obtain benefits and can also prevent vehicles from obtaining the right-of-way priority through false information.Compared with the virtual traffic light with a fixed threshold of the number of green lights, the virtual traffic lights protected by the group strategy show some improvement in the overall average driving time and the average driving time of high-value vehicles.

Key words: V2V environment, virtual traffic light, cost sharing cooperative game, group strategy-proof auction mechanism, Shapley value, average duration time

中图分类号: