The main objective of this study was to develop a new driving assistance system that could help less experienced drivers improve their driving skills. A statistical method is described that was developed to extract distinctions between experienced and less experienced drivers. This paper makes three key contributions. The first involves a technology for feature extraction based on AdaBoost, which selects a small number of features critical for operation between experienced and less experienced drivers. The second involves a simple definition for experienced and less experienced drivers. The third involves the introduction of wavelet transforms that were used to analyze the frequency characteristics of driver operations. A series of experiments was performed using a driving simulator on a specially designed course that included several curves and then used the proposed method to extract features of driving operations that demonstrated the differences between the two groups. (Author/publisher)
Abstract