Proposal of geographic information systems methodology for quality control procedures of data obtained in naturalistic driving studies

Abstract

The primary goal of naturalistic driving studies is to provide a comprehensive observation of the driver's behaviour under real-life conditions by measureing a great number of parameters at high temporal frequencies. Achieving this goal, however, is a complex endeavor that faces many challenges such as the complexity of the vehicle instrumentation during the phase of data collection, and the difficult handling of large data volumes during the phase of data analysis. These drawbacks often cause episodes of data losses. Improving the technical aspects of the collection of naturalistic data is of paramount importance to increase the return of the investment made in it. An aspect to consider is the control of the quality of data obtained. The procedures commonly applied often present a high workload and low levels of reliability in the results. For this reason, new and more efficient procedures should be implemented. This study proposes an innovative methodology for the quality control of these data through geographic information systems (GIS), which permit a fast, efficient and effective data evaluation. Relative and absolute errors can be located through cartographic representation obtained with GIS. Some graphical examples about the detection of the errors are shown.