时变的精准的城市居民OD出行预测，是实施城市交通精细化管理的前提。本项目针对环境和关键设施对出行影响显著、出行空间效应具有时变特征的实际，基于融合的多源异构交通大数据，分析出行时空特征及其影响因素；以出行时间可靠性为基础，构建时变的空间权重矩阵生成模型；建立面向环境和设施等级影响的OD出行时变空间自相关模型，揭示出行多因素影响机理，科学预测城市OD出行时空演化规律。本项目构建的时变OD空间自相关模型，克服现有研究将空间权重矩阵外生固化的局限性，将空间自相关分为起点自相关、终点自相关和OD自相关，能更准确分析各因素对出行的方向和强度的影响，并深刻揭示环境和设施等因素对出行的影响机理及空间效应。该模型具有可扩展性，可应用到其他交通OD空间自相关问题。项目成果学术上将完善和发展城市出行行为理论和方法，为城市交通规划、建设和管理提供决策支持。

The precise time-varying forecasting of urban residents' travel OD is the prerequisite for the implementation of the fine management of urban traffic. This project focuses on the facts that the environment and key facilities have significant impacts on travel and the spatial effect of travel has time-varying characteristics as well, and analyzes spatio-temporal features of travel and its influence factors based on the fusion of multi-source transportation big data. The generative model of the time-varying spatial weight matrix is constructed on the basis of travel time reliability. The time-varying spatial autocorrelation model of travel OD matrix is established for the impacts of the environment and facility level, in order to reveal the multi-factor influence mechanism of travel, and scientifically predict the spatio-temporal evolution of urban travel OD. The established time-varying OD spatial autocorrelation model by this project will overcome the limitations of the spatial weight matrix exogenesis in existing literatures. Dividing the influence of spatial dependence into the origin dependence, destination dependence, and OD dependence, can more accurately analyze the multi-factor influences on the travel direction and strength, and deeply reveal the travel influence mechanism and spatial effects of factors such as the environment and facilities. The model is extensible and can be applied to other travel OD spatial autocorrelation problems. The results of the project will improve and develop the theory and methodologies of urban travel behavior, and provide decision-making support for urban transportation planning, construction, and management.