Driver source prediction and visualization system based on mobile phone data

Publication
Undergraduate Thesis

With the development of urbanization never seen before, the condition of traffic congestion, transit service quality and mergency response capability become more severe. Existing research and engineering practice tend to focus on the analysis and control of phenomena, but lack of probing the underlying causes of the problem.The rise of big data and the proposed source prediction methods provide an important foundation for analysis of the traffic problems’ underlying causes and solving the traffic problems from the source. The project combines the idea of traffic engineering, with the technology of software engineering, uncovers the mobility of residents from large-scale mobile phone data, extracts the trip information and calculating the O-D matrix, with the residents’ trip paths predicted by traffic assignment, forecasts the driver source of city road network and the passenger source of block section between station, locate congested road clusters, establishes an interactive visual analytics system.

The system workflow consists of two phase: preprocessing and visualization. Constructing the three subsystems: driver source prediction, passenger source prediction and congested cluster analysis, provides the visual charts, particle trajectories view , map zoom , screen shots, custom maps and so on. While the system uses asynchronous loading , offline maps, vector data compression techniques to provide the high-quality and efficient services for users. Considering the portability ,the system supports multiple operating systems.

The system will comprehensively consider the scientificity and practical basis of source prediction service, explicitly presenting the bipartite relations between traffic resources and residents to traffic planners and researchers by closely combining the user’s actual needs. Through mastery of the main sources of traffic problem, department of transportation could take appropriate measures, such as traffic dispersion and traffic restriction, to ease the traffic congestion, improving the quality of service and emergency response capability. Therefore, the presented project has a good prospect and practical value.

Outstanding Undergraduate Thesis (2nd) (Top 1%) - Central South University - 2014

Advisor: Pu Wang, Lin Jiang, Wei Liu