Studying the influence of cognitive load on driver's performances by a Fuzzy analysis of Lane Keeping in a drive simulation.

Abstract

The psycho-physical state of a driver has been widely recognized as the crucial point in any issue concerning the development of models headed to improve the vehicle safety, either inherent and active, so much so that almost all the new in-vehicle technology, currently developing at a rapid rate, introduces devices to continuously monitor the driver. This paper describes the architecture of a real time, performance-based, driver monitoring system able to detect the decrease in driver performances due to driver distraction, fatigue, sleepiness and alcohol or drugs ingestion. The system processes the instantaneous lateral position of the vehicle on the road. This allows to work out an index of the lane keeping precision by means of the lateral position standard deviation (SDLP). This latter and the road environment complexity has been processed by a fuzzy inference system that has, as an output, a score reflecting the driver's ability to maintain adequate lane-tracking movements for a given road scenario. Fuzzy membership functions and inference rules has been based and optimized on data obtained on 12 subjects performing driving simulation under both baseline condition and two different cognitive overload situations induced by different secondary tasks, one with visual distraction, the other characterized by a pure cognitive load, respectively. Aim of the work is to attain to a black-box type devices that could both provide warnings or reminding in case of risky driving and encourage the driver to improve his behavior. Advantages would also come for parents of novice drivers promptly alerted for improper driving and even for the car insurance companies that could reward safe drivers.