一種新的基於高斯混合模型之行為塑模方法用於智慧型手錶之駕駛者識別;A Novel GMM-Based Behavioral Modeling Approach for Smartwatch-Based Driver Recognition

NCUIR > College of Electrical Engineering & Computer Science > Graduate Institute of Computer Science and Information Engineering > Electronic Thesis & Dissertation > Item 987654321/77499

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/77499

Title:	一種新的基於高斯混合模型之行為塑模方法用於智慧型手錶之駕駛者識別;A Novel GMM-Based Behavioral Modeling Approach for Smartwatch-Based Driver Recognition
Authors:	楊青翰;Yang, Ching-Han
Contributors:	資訊工程學系
Keywords:	加速感測器;駕駛者認證;駕駛者識別;高斯混合模型;方位感測器;智慧型手錶;Accelerometer sensor;Driver authentication;Driver identification;Gaussian mixture model;Orientation sensor;Smartwatch
Date:	2018-06-25
Issue Date:	2018-08-31 14:46:02 (UTC+8)
Publisher:	國立中央大學
Abstract:	每位駕駛者都有屬於他們獨特的駕駛習慣，且通常在不同駕駛情境下，同一位駕駛者握住或操作方向盤的方式也不同。在本研究中，我們提出一種新的基於高斯混合模型塑模方法，此塑模方法可以改善傳統高斯混合模型在駕駛行為塑模上的問題。此外，我們提出的塑模方法可以應用在建構較佳的智慧型手錶感測器(如加速度計、方位感測器)之駕駛者認證系統或駕駛者識別系統上。為了驗證我們提出的方法之可行性，我們建構兩個實驗系統，分別為駕駛者認證系統與駕駛者識別系統。對於駕駛者認證系統的實驗結果顯示，在模擬環境中相等錯誤率(Equal Error Rate)可達4.46%；在真實駕駛環境中相等錯誤率達可11.35%。對於駕駛者識別系統而言，實驗結果顯示，在模擬環境中識別率可達87.16%；在真實環境中識別率可達73.07%。上述實驗結果皆比傳統高斯混合模型方法佳，因此，可以證實我們提出新的基於高斯混合模型塑模方法具有可用性。;All drivers have their own distinct driving habits, and usually hold and operate the steering wheel differently in different driving scenarios. In this study, we proposed a novel Gaussian mixture model (GMM)-based method that can improve the traditional GMM in modeling driving behavior. This new method can be applied to build a better driver authentication or identification system based on the accelerometer and orientation sensor of a smartwatch. To demonstrate the feasibility of the proposed method, we created two experimental systems that analyzes driving behavior using the built-in sensors of a smartwatch. The experimental results for driver authentication—an equal error rate (EER) of 4.46% in the simulated environment and an EER of 11.35% in the real-driving environment—confirm the feasibility of this approach. For driver identification, the experimental results indicated that the proposed approach had identification rates of 87.16% in a simulated environment and 73.07% in a real-driving environment.
Appears in Collections:	[Graduate Institute of Computer Science and Information Engineering] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	97	View/Open

社群 sharing

Loading...