Understanding the travel flux between urban blocks is fundamental for traffic demand prediction, urban area planning and urban traffic management. However, the uncertainty of human mobility patterns and the complexity of urban transportation systems usually yield challenges in accurately estimating the travel flux within a city. Thus, we propose a novel travel flux estimation method that integrates traffic flow characteristics (traffic volume and travel time), spatio-temporal autocorrelation, and travel purpose correlation. First, the geographically weighted method was used to model and verify the spatio-temporal autocorrelation of origin–destination flows, whereas the purpose correlation of origin–destination flows was expressed through the function feature vectors of urban blocks. Then, a multi-objective bi-level programming model, according to the generalized least squares method, was constructed to estimate the travel flux between blocks. This was used to solve the problem of accurate estimation of travel flux by combining microscopic traffic flow characteristics with macroscopic spatio-temporal and purpose characteristics. Finally, an empirical analysis of the Hankou district, Wuhan City, demonstrated that in contrast to the existing method, the accuracy of the proposed method for predicting the human travel flux improved by approximately 20%. The estimated results were consistent with the spatial distribution pattern of human travel. Moreover, these results can provide targeted decision support for planning urban spaces, allocating urban resources, and guiding vehicular traffic.

中文翻译：

---

结合时空和目的相关性估计城市街区之间的出行流量

了解城市街区之间的出行流量对于交通需求预测、城市区域规划和城市交通管理至关重要。然而，人类流动模式的不确定性和城市交通系统的复杂性通常会给准确估计城市内的出行流量带来挑战。因此，我们提出了一种新颖的出行流量估计方法，该方法集成了交通流特征（交通量和出行时间）、时空自相关和出行目的相关性。首先，采用地理加权方法对始发地-目的地流的时空自相关性进行建模和验证，而始发地-目的地流的目的相关性则通过城市街区的功能特征向量来表达。然后，根据广义最小二乘法，构建了多目标双层规划模型来估计街区间的出行流量。用于将微观交通流特征与宏观时空、目的特征相结合，解决出行流量的准确估算问题。最后，对武汉市汉口区的实证分析表明，与现有方法相比，该方法预测人类出行流量的准确率提高了约20%。估算结果与人类出行的空间分布格局一致。此外，这些结果还可以为规划城市空间、配置城市资源、引导车辆交通提供有针对性的决策支持。