PROBLEM TO BE SOLVED: To provide a road surface state discrimination method which improves a calculation speed by performing appropriate path restriction and further improves discrimination accuracy of a road surface state.SOLUTION: When a time-series waveform of a tire vibration detected by an acceleration sensor is windowed with a time T by a window means, the time-series waveform of the tire vibration at every time window is extracted, a feature vector Xat every time window is calculated and, thereafter, a kernel function K(X,Y) is calculated from the feature vector Xat every time window and a road surface feature vector Yas the feature vector at every time window calculated from the time-series waveform of the tire vibration determined at every road surface state calculated beforehand, each of paths is weighed by a weighing parameter ω(i,j) using a time-series waveform length correspondence value on the time-series waveform of a tire vibration upon measurement and the time-series waveform length correspondence value on the time-series waveform of a tire vibration determined beforehand and, then, the kernel function is calculated.SELECTED DRAWING: Figure 10
【課題】適切なパス制限を行うことで、計算速度を向上させるとともに、路面状態の判別精度を更に向上させる。【解決手段】加速度センサーにより検出したタイヤ振動の時系列波形を窓掛け手段により時間Tで窓掛けして、時間窓毎のタイヤ振動の時系列波形を抽出して時間窓毎の特徴ベクトルXiを算出した後、時間窓毎の特徴ベクトルXiと、予め算出しておいた路面状態毎に求められたタイヤ振動の時系列波形から算出された時間窓毎の特徴ベクトルである路面特徴ベクトルYjとからカーネル関数KA(X,Y)を算出する際、パスのそれぞれに、計測時のタイヤ振動の時系列波形における時系列波形長さ対応値と、予め求めておいたタイヤ振動の時系列波形における時系列波形長さ対応値とを用いた重み付けパラメータω(i,j)により重み付けしてから、カーネル関数を算出するようにした。【選択図】図10
ROAD SURFACE STATE DISCRIMINATION METHOD
路面状態判別方法
2016-06-20
Patent
Elektronische Ressource
Japanisch
ROAD SURFACE STATE DISCRIMINATION METHOD, AND ROAD SURFACE STATE DISCRIMINATION DEVICE
Europäisches Patentamt | 2018
|ROAD STATE DISCRIMINATION METHOD AND ROAD STATE DISCRIMINATION DEVICE
Europäisches Patentamt | 2019
|ROAD FACE STATE DISCRIMINATION METHOD AND ROAD FACE STATE DISCRIMINATION DEVICE
Europäisches Patentamt | 2019
|