Potential of Steering Wheel Information to Detect Driver Drowsiness andAssociated Lane Departure

摘要

The report summarizes research to date concerned with identification of drowsydrivers from information contained in steering wheel and lane position data. This report concludes that the type of steering wheel sensor can influence some of the more popular statistics promoted for drowsy driver identification; weather conditions have a strong influence on the same statistics, and stochastic characterizations of steering wheel and lane position processes appear to have strong similarities. One goal of the research was to assess the potential of the joint stochastic information contained in steering wheel and lane position data of identifying trends in drowsiness and for predicting imminent lane departure associated with the same. A second goal concerned identification of hardware and software issues which could improve that system's performance. A third goal was to investigate the potential value of the stochastic structure of lane position and steering wheel position data, relative to the more commonly studied summary statistics.