METHOD AND DEVICE FOR DETERMINING STATE OF ROAD SURFACE

A time-series waveform of tire vibration detected by an acceleration sensor is windowed by a windowing means, time-series waveforms of the tire vibration are extracted for respective time windows, and feature vectors of the respective time windows are calculated. Then kernel functions are calculated from the feature vectors of the respective time windows and road surface feature vectors, which are the feature vectors for the respective time windows calculated from the time-series waveform of the tire vibration obtained for distinctive road surface conditions calculated in advance. And the road surface condition is determined by comparing values of discriminant functions using the kernel functions. As a result, the road surface condition can be determined from the time-series waveform of the tire vibration without detecting peak positions or measuring the wheel speed. Moreover, robustness against changes in tire size can be added to the determination of the road surface condition.