Better traffic estimation can be achieved by integrating multiple data sources. However, it is not an easy task due to many issues such as differences in formats, spatio-temporal resolutions, availability and reliability. In this study, we developed an incremental Unscented Kalman Filter (UKF) to effectively deal with data from multiple sources for a real-time motorway traffic estimation problem. The estimates produced by our model were compared with those from the incremental Extended Kalman Filter (EKF). The results showed similar performance between the incremental UKF and the incremental EKF, but our proposed framework proved to be more reliable due to smaller variance estimates, particularly during free-flow periods. The framework was also applied to estimate flow and speed in cases where data were incomplete. It has been shown that by combining multiple data sources, the filter can compensate for the deficiency of each source to produce more accurate estimates.

通过集成多个数据源可以实现更好的流量估计。然而，由于格式差异、时空分辨率、可用性和可靠性等诸多问题，这并不是一件容易的事。在这项研究中，我们开发了一种增量无迹卡尔曼滤波器 (UKF)，以有效处理来自多个来源的数据，以解决实时高速公路交通估计问题。我们的模型产生的估计值与增量扩展卡尔曼滤波器（EKF）的估计值进行了比较。结果显示增量 UKF 和增量 EKF 之间的性能相似，但我们提出的框架被证明是更可靠的，因为较小的方差估计，特别是在自由流动期间。该框架还用于在数据不完整的情况下估计流量和速度。