Turning movements are one of the key inputs required for several traffic studies. Several methods have been developed to measure them. However, present techniques have high operational or capital costs, which motivate researchers to develop new techniques to estimate turning movements. However, there is neither a flexible technique available to make best use of different available information types, nor a framework that supports deciding additional data to achieve a target accuracy. This paper proposes a new methodology using all available data to identify the subspace containing all solutions and determine its centroid; thus, providing the most realistic and non-extreme solution. In addition, a framework, including scenarios with different data combinations, is developed with capability to evaluate the proposed solution and then locate further measurements to achieve the target accuracy. The framework is validated using a considerable set of intersections at Edmonton city, Canada. The results show that the proposed framework can achieve the target accuracy with minimum field measurements saving time, effort and cost.

转弯运动是多项交通研究所需的关键输入之一。已经开发了几种方法来测量它们。然而，目前的技术具有很高的运营或资本成本，这促使研究人员开发新技术来估计转弯运动。然而，既没有一种灵活的技术可以充分利用不同的可用信息类型，也没有一种框架来支持决定额外的数据以实现目标精度。本文提出了一种使用所有可用数据来识别包含所有解的子空间并确定其质心的新方法；因此，提供了最现实和非极端的解决方案。另外一个框架，包括不同数据组合的场景，开发具有评估建议解决方案的能力，然后定位进一步的测量以实现目标精度。该框架使用加拿大埃德蒙顿市的大量交叉路口进行了验证。结果表明，所提出的框架可以以最少的现场测量达到目标精度，从而节省了时间、精力和成本。