Abstract
This paper tries to develop a method to detect driver's specific actions in a non-intrusive manner. The Higher-order Local Auto-Correlation (HLAC) feature is extracted from the data of a pressure distribution on the seat cushion measured with a tactile sheet, and a linear discriminant analysis is performed to classify the data. The results suggest that the method may be successfully applicable to both evaluation of driver's ''preparedness'' to depress the brake pedal and detection of driver's actions that are unnecessary for driving.
