Trace-based analysis and spatio-temporal modelling of human mobility considering geographic restrictions for the performance evaluation of wireless communication networks

对人的流动性进行建模是几个科学领域的相关课题。在计算机科学中，由于其对无线通信网络中的性能评估的重大影响而引起人们的高度兴趣。因此，已经在移动性建模的背景下进行了重要的研究。目前，智能手机和平板电脑等移动设备的普及率稳步上升，再加上所包含的定位和通信技术，使得能够有效地跟踪人体轨迹。这种轨迹的足够大的数据基础(所谓的“轨迹”)可用于分析基本的机动性特征。这导致了过去几年在人类流动性模型领域的重要进展。然而，项目调查者之前的工作表明，在人类流动性模型领域仍然存在显著的研究空白。一方面，目前的流动性模型缺乏与其他痕迹的验证。另一方面，到目前为止使用的跟踪大多过于粗粒度、特定于场景，或者基于小样本大小。此外，目前的痕迹分析和模型是基于假设的，如起点和目的地之间的航班，这在人员流动的背景下是不现实的。在轨迹分析或基于轨迹的建模过程中，通常不考虑地理限制(例如，像OpenStreetMap这样的街道地图)。然而，可以预见的是，地理限制在很大程度上影响了特征特征的分析和建模。这些研究差距挑战了当前痕迹分析和人员流动性建模结果的真实性和可信度，以及基于这些模型的性能评估结果。本项目的目标是基于新的痕迹基础建立一个现实的和有效的人类流动性建模。应用的重点是无线通信网络的性能评估。然而，由痕迹分析得出的人类流动性的特征也可以应用于其他领域，因为它们的基本性质。这些领域包括交通系统规划、旅行需求预测、生物和无线病毒传播等。

Modeling of human mobility is a relevant topic in several scientific areas. In computer science, it is of high interest, e.g., due to its significant impact on performance evaluation in wireless communication networks. Thus, significant research has already been conducted in the context of mobility modeling. The steadily rising prevalence of mobile devices, e.g., smart phones and tablets, combined with the included positioning and communication technology nowadays enable the effective tracking of human trajectories. A sufficiently large data basis of such trajectories (so-called "traces") can be used for analyzing fundamental mobility characteristics. This has led to important advances in the area of human mobility models during the last years.However, previous work by the project investigator has shown that there are still significant research gaps in the field of human mobility modeling. On the one hand, current mobility models lack a validation with other traces. On the other hand, the traces used so far are mostly too coarse-grained, scenario-specific, or are based on a small sample size. Furthermore, current trace analyses and models are based on assumptions, like flights between start and destination, which are unrealistic in the context of human mobility. Geographic restrictions (e.g. street maps like OpenStreetMap) are usually not considered during trace analysis or trace-based modeling. However, it is to be expected that geographic restrictions impact the analysis of characteristic features as well as modeling at a high degree. These research gaps challenge the realism and credibility of current results from trace analysis and human mobility modeling as well as the results of the performance evaluation based on these models.This project's goal is a realistic and validated modeling for human mobility based on the new trace basis. The application focus is on the performance evaluation in wireless communication networks. Nevertheless, the characteristic features of human mobility yielded by the trace analysis can also be applied to other domains due to their fundamental nature. These domains are transportation system planning, travel demand forecasting, and biological and wireless virus spreading, among others.