基于心理学和经济学理论的"人本自治交通"科学假想及群组动画推演验证

This proposal, for the first time, proposes a "Human-initiative autonomous traffic regulation"theory, inspired by psychology and economics principles. It is based on the following key assumptions: (1) Each car will be installed with an E-eye device that can be used to monitor other cars on the road, which enables each car to play the role of law enforcement and the role by law enforcement. (2) Each car will win certain amount of rewards by reporting the illegal driving behavior of other cars via its installed E-eye system. (3) A car's traffic avoidance behavior will be given certain amount of rewards if it is identified and reported by other car's E-eye system. (4) In order to drive on the road, each car need to pay traffic resource cost. In this way, every car owner will make the most rational decision to maximize the utility of his/her own traffic resource. (5) Everyone in the society will be evenly allocated a certain amount of traffic resource regularly, without charge. An interdisciplinary research team, combining theories and methodologies from information science, psychology, and economics, will develop a tractable computational model to quantitatively model the driver behavior function in psychological space, where each driver-car is studied as a basic micro-unit. The pipeline of our approach consists of the following key steps: proposing scientific assumptions -> construction of computational model -> ground-truth data collection -> model-based simulation -> simulation based validation and evaluation. Finally, we will systematically model the dynamic evolving of the traffic after our proposed computational model is incorporated, and further validate the theoretical feasibility of the proposed computational model based on data-driven traffic simulation.

根据心理学与经济学理论，创新地提出"人本自治交通"科学假想：（1）对每辆车安装可监控其它车辆交通行为的"电子眼"设备，实现车辆之间的相互监控；（2）每个驾驶人都可通过举报其它车辆的违章行为获得一定的奖励，由违章驾驶人支付奖励；（3）车辆的主动避让行为能被"电子眼"识别，并会获得一定奖励；（4）收取交通资源使用费，促使所有出行人根据其购买力理性选择交通方式，实现有限交通资源的效用最大化；（5）定期向所有驾驶人平均发放一定量的交通资源免费使用卷，用以保证公平；（6）违章驾驶行为将被扣除一定的驾驶行为信用值，信用值只能通过系统对所有驾驶人定期平均发放获得，其使用具有时效性，一定时间内每个驾驶人的信用值是有限的，扣除完后驾驶人将被终止驾驶，用以约束违章行为。据此对实施该假想后的车辆流动演化过程进行多智能体系统建模，采用群组动画技术推演验证假想的理论可行性，为交通治理提供新的研究思路和方法手段。

本项目经过4年的深入探索研究，应用心理学与经济学的相关理论，提出了"人本自治交通"科学假想：（1）对每辆车安装可监控其它车辆交通行为的"电子眼"设备，实现车辆之间的相互监控；（2）每个驾驶人都可通过举报其它车辆的违章行为获得一定的奖励，由违章驾驶人支付奖励；（3）车辆的主动避让行为能被"电子眼"识别，并会获得一定奖励；（4）收取交通资源使用费，促使所有出行人根据其购买力理性选择交通方式，实现有限交通资源的效用最大化；（5）定期向所有驾驶人平均发放一定量的交通资源免费使用卷，用以保证公平；（6）违章驾驶行为将被扣除一定的驾驶行为信用值，信用值只能通过系统对所有驾驶人定期平均发放获得，其使用具有时效性，一定时间内每个驾驶人的信用值是有限的，扣除完后驾驶人将被终止驾驶，用以约束违章行为。我们构建了一个由信息学科、心理学科、经济学科等多学科参与的交叉研究团队，按照“提出科学假想，构造计算模型，采集实验数据，基于模型仿真，仿真验证评估”的研究步骤开展项目。进而，对基于该假想的车辆流动演化过程进行多智能体系统建模，并采用群组动画技术推演验证了假想的理论可行性，为交通治理提供了新的研究思路和方法手段。目前，共计发表论文41篇，其中在ACM TOG、IEEE TPAMI、IEEE TIP、IEEE TCSVT、IEEE TSMC: System等SCI期刊发表论文28篇。