汽車與日常生活密不可分，而車輛安全的問題卻也持續發生中，而隨著生物辨識技術的發展，駕駛者識別方法也愈來愈多樣，雖然有助於減少車輛安全問題，但還是有缺失或是未發展完全。 隨著穿戴型裝置技術成熟與目前汽車大廠的軟體應用發展趨勢，未來，汽車與智慧型手錶結合應用將蓬勃發展。目前市面上的智慧型手錶大多內建多種感測器，如加速度計、陀螺儀、磁力計與方位感測器等，這些感測器為駕駛者生物特徵識別技術帶來新的可能性。 在本研究中，將利用加速度計與方位感測器資料作為生物特徵，提出一個高斯混合模型與支持向量機結合的駕駛者驗證方法。而為了評估此駕駛者驗證方法的效能，邀請50位參與者收集模擬環境的駕駛行為資料，並且進行效能評估實驗。實驗結果顯示駕駛者驗證的等誤差率為7.63%。我們認為在此方法上，可以進行一些改善以獲得更好結果。 ;Today, vehicles have been an essential part of our daily life. One-third of drivers admit they have left their vehicle while it is idling, which makes the vehicle an easy target of theft. In recent years, many verification methods had been developed, but there is still a room for better result. In this research, a novel method of driver verification by combining Gaussian Mixture Model (GMM) and Support Vector Machine (SVM) is proposed. The proposed method is based on the hypothesis that drivers have their own specific driving behaviors; and the driving behaviors can be captured from smartwatch sensors and used as behavioral biometrics for driver recognition. In order to validate this hypothesis, a simulation system has been established to collect 50 drivers’ driving behavioral information, and the experimental result shows there are same methods to improve this experimental approach.
