A Micro-Mobility Analysis System (μMAS) is developed to characterize drivers based on their driving habits. This is achieved by solving a pattern recognition problem, which can be divided into four phases: data collection, data pre-processing, parameter extraction, and classification. These phases are discussed in detail and the habits used to distinguish between drivers are various parameters associated with their general characteristics, turning behavior, and lane change mentality. A cross-validation simulation is implemented to gauge the performance of μMAS. The results indicate that the system was successful in identifying one of the three drivers, but not the other two. The various factors contributing to this performance are discussed. The techniques developed in this study can be used to measure distance between drivers and place them into clusters, which can then be used to assess whether they drive in a safe or unsafe manner. 

Cite and download the thesis:

S. Mukherji, "Development of a Micro-Mobility Analysis System Using Precise GPS Traces," Undergraduate Honors Thesis, University of Texas at Austin, December 2014.