自动驾驶对交通运输系统规划的影响综述

车辆是交通运输系统的重要组成部分，伴随着自动驾驶技术的发展与应用，交通运输系统 将发生深刻变革。本文聚焦于自动驾驶技术对交通运输系统规划的影响，综述自动驾驶特点下 交通数据采集与管理手段、土地利用、停车需求、交通供需、交通需求预测、交通网络布局等方面 发生的新变化。在此基础上，从规划的角度出发，实现对自动驾驶环境下交通运输系统的再认 知，总结自动驾驶环境下交通需求预测、城市交通网络布局等新的交通规划方法与技术。通过对 交通运输系统的再认知发现：在自动驾驶环境下，交通数据具有细粒度、高鲜度的新特点；土地利 用模式将发生改变，城市将呈扩张和去工业化趋势，停车需求减小；交通系统供给能力和可靠性 提高，出行需求的时空分布将更为分散。交通系统规划方法的变化体现在交通需求预测和交通 网络布局两个层面：交通需求预测框架从“四步”框架转变为模型组合化和出行行为一体化的预 测框架，同时，需求预测的各阶段需引入对自动驾驶特征及其系统性影响的分析；交通网络布局 设计采用连续时域上的布局设计框架，有望解决传统交通网络布局设计的时滞性问题，可适应并 服务于动态变化的土地利用及交通需求。本研究认为，未来需重点研究自动驾驶对交通安全、交 通拥堵、公共交通规划、慢行交通规划等方面的影响。此外，解决自动驾驶实测数据缺乏的困境、 解析异构交通阶段交通系统的运作机理、应对交通需求反弹引起的供不应求、评估难以衡量的外 部成本等问题将是未来研究的难点。

A Review of the Impact of Autonomous Driving on Transportation Planning

As vehicle is an important component of the transportation system, the development and application of autonomous driving is triggering a revolution of the transportation system. This paper focuses on the impact of autonomous driving on transportation planning. By summarizing the technologies, status in quo and prospect of transportation planning, this paper reviews the revolution in data acquisition and management, land use, parking demand, supply-demand analysis, traffic prediction and transportation network design in the environment of autonomous driving. Furthermore, from the perspective of transportation planning, this paper refreshes the understanding of transportation system with consideration of autonomous driving. This paper also proposes novel philosophy and methods of transportation planning, which provide a new analysis framework and research methods for traffic demand forecasting and traffic network design in the environment of autonomous driving. The understanding of transportation system could be refreshed from three aspects. First, traffic data will be more fine-grained and fresh.Second, changes in land use patterns will cause cities to expand and de-industrialize, and the demand for parking will decrease. Third, the supply capacity and reliability of the transportation system will be improved, and greater dispersion will take place in the temporal and spatial distribution of travel demand. Changes in methodologies of the transportation system planning are reflected in two aspects: demand forecasting and traffic network design. First, the framework of demand forecasting will be transformed from a four-step framework to a framework of model combination and travel behavior integration. In addition, in each step of demand forecasting, the characteristics of autonomous driving and its systematic impact should be analyzed. Second, the traffic network design will adopt a continuous-time dependent design framework, which is expected to improve traditional network design by solving the issue of responsive delay. This framework can adapt to and serve the dynamic land use and traffic demand. This study suggests that future research should devote the major efforts to investigating the impact of autonomous driving on traffic safety, congestion, public transit planning and non-motorized transportation planning. In addition, the research difficulties will lay on the following aspects: solving the issue of lacking real-world data of autonomous driving; revealing the mechanism of the transportation operation in the heterogeneous-traffic stage; coping with the situation when demand exceeds supply due to the traffic demand rebound, and evaluating the external costs which are difficult to measure.