Accurately estimating traffic variables across unequipped portions of a network remains a significant challenge due to the limited coverage of sensor-equipped links, such as loop detectors and probe vehicles. A common approach is to apply uniform scaling, treating unequipped links as equivalent to equipped ones. This study introduces a novel framework to improve traffic variable estimation by integrating statistical scaling methods with geospatial imputation techniques. Two main approaches are proposed: (1) Statistical Scaling, which includes hierarchical and non-hierarchical network approaches, and (2) Geospatial Imputation, based on variogram modeling. The hierarchical scaling method categorizes the network into several levels according to spatial and functional characteristics, applying tailored scaling factors to each category. In contrast, the non-hierarchical method uses a uniform scaling factor across all links, ignoring network heterogeneity. The variogram-based geospatial imputation leverages spatial correlations to estimate traffic variables for unequipped links, capturing spatial dependencies in urban road networks. Validation results indicate that the hierarchical scaling approach provides the most accurate estimates, achieving reliable performance even with as low as 5% uniform detector coverage. Although the variogram-based method yields strong results, it is slightly less effective than the hierarchical scaling approach but outperforms the non-hierarchical method.

翻译：由于配备传感器（如环形检测器和浮动车）的路段覆盖范围有限，准确估计路网中未配备传感器部分的交通变量仍然是一个重大挑战。一种常见的方法是应用统一尺度缩放，将未配备传感器的路段视为与已配备路段等同。本研究引入了一个新颖的框架，通过将统计尺度缩放方法与地理空间插补技术相结合，以改进交通变量的估计。提出了两种主要方法：(1) 统计尺度缩放，包括分层和非分层的网络方法；(2) 基于变异函数建模的地理空间插补。分层尺度缩放方法根据空间和功能特征将路网划分为多个层级，并对每个类别应用定制的缩放因子。相比之下，非分层方法在所有路段上使用统一的缩放因子，忽略了网络的异质性。基于变异函数的地理空间插补利用空间相关性来估计未配备传感器路段的交通变量，捕捉了城市道路网络中的空间依赖性。验证结果表明，分层尺度缩放方法提供了最准确的估计，即使在均匀检测器覆盖率低至5%的情况下也能实现可靠的性能。尽管基于变异函数的方法产生了良好的结果，但其效果略逊于分层尺度缩放方法，但优于非分层方法。