本研究提出基於路況訊息分級之個人化駕駛輔助機制， 能夠針對不同道路特徵進行量化分 析，並以模糊邏輯概念實現道路等級設計，提供駕駛以直覺的方式，在短時間內了解路況的呈現 或變化。此外利用類神經網路建立駕駛行為模型， 可分析駕駛在不同道路特徵下的典型駕駛行 為，並建立駕駛可容忍之道路等級。此建模之輸出結果，可做為道路訊息過濾機制的判斷條件， 並透過所開發之駕駛模擬平台，分別比較傳統警示方式與個人化警示方式，可驗證我們所提出之 輔助方法， 有效降低警示系統對於駕駛者不必要之警示干擾。

This study presents the personal driving assistance mechanism based on road information grade. It can provide quantitative analysis in various road characteristics, and realize the road information grade design by the concept of fuzzy logic. The proposed approach provides drivers an intuitive way to understand rapidly the variation of the road condition. Moreover, the neural network is used to construct the driving behavior model. It can be utilized to analyze the typical driving behavior in various road characteristics and establish the road level which driver can tolerance. The output of driving behavior model can be regarded as determined condition for road information filtering. In addition, the driving simulation platform is developed to compare the traditional and personal alert way. It can verify that our proposed assistance system is able to reduce the non-necessary alert which may cause the driver interference.