PROJECT TITLE

Mobility Data Mining for Intelligent Transportation

PRIMARY INVESTIGATOR

Yi Liu Murphey, PhD
Professor and Chair, Electrical and Computer Engineering; Director, Intelligent Systems Lab, U-M Dearborn

PROJECT DATES

May 2014 – April 2016

PROJECT TEAM

William Grosky, PhD
Professor and Chair, Computer and Information Science, U-M Dearborn

Bochen Jia, PhD
Assistant Professor, Industrial and Manufacturing Systems Engineering, U-M Dearborn

Yung-Wen Liu, PhD
Associate Professor, Industrial and Manufacturing Systems Engineering, U-M Dearborn

OBJECTIVE

To develop effective mobility data mining techniques and intelligent transportation systems that can be used to improve driving safety through V2V and V2I communication systems.

APPROACH

Data from vehicles and the real-time traffic information from an information service provider, such as Nokia/Navteq, are transmitted to the Mobility Data Mining Center (MDMC). The MDMC will predict each driver’s traveling speed and time, which are then sent back to the respective vehicles and the Traffic Analytics and Control Centers (TACC) for traffic management and emissions management. Furthermore, the MDMC generates mobility data and patterns, and traffic congestion forecasts, which are communicated to the TACC.

OUTCOMES

GPS errors can cause a serious problem for all vehicle safety systems based on V2V. Our research discovers that GPS errors can cause vehicle trajectories being mapped to wrong road sections, inaccurate detection of pre-crash warnings from approaching remote vehicles, and so forth.

For calculating the distance between two GPS locations, we recommend using a geodistance function based on the Haversine formula, which is more accurate and computationally efficient than other well-known geodistance functions.


CATEGORY: BIG DATA
|