Date of Award

12-14-2015

Document Type

Open Access Thesis

Department

Civil and Environmental Engineering

First Advisor

Nathan N. Huynh

Abstract

Transportation apps are playing a positive role for today’s technology-driven users. They provide users with a convenient and flexible tool to access transportation data and services, as well as collect and manage data. In many of these apps, such as Google Maps, their operations rely on the effectiveness of the voice recognition system. For the existing and new apps to be truly effective, the built-in voice recognition system needs to be robust (i.e., being able to recognize words spoken in different pitch and tone). The goal of this study is to assess three post-processing classifiers (i.e., bag-of-sentences, support vector machine, and maximum entropy) to enhance the commonly used Google’s voice recognition system. The experiments investigated three factors (original phrasing, reduced phrasing, and personalized phrasing) at three levels (zero training repetition, 5 training repetitions, and 10 training repetitions). Results indicated that personal phrasing yielded the highest correctness and that training the device to recognize an individual’s voice improved correctness as well. Although simplistic, the bag-of-sentences classifier significantly improved voice recognition correctness. The classification efficiency of the maximum entropy and support vector machine algorithms was found to be nearly identical. These results suggest that post-processing techniques could significantly enhance Google’s voice recognition system.

Rights

© 2015, Md Majbah Uddin

Share

COinS