Abstract

The goal of this thesis is to develop a voice-controlled human-robot interface (HRI) which allows a person to control and communicate with a robot. Dragon NaturallySpeaking, a commercially available automatic speech recognition engine, was chosen for the development of the proposed HRI. In order to achieve the goal, the Dragon software is used to create custom commands (or macros) which must satisfy the tasks of (a) directly controlling the robot with voice, (b) writing a robot program with voice, and (c) developing a HRI which allows the human and robot to communicate with each other using speech. The key is to generate keystrokes upon recognizing the speech and three types of macro including step-by-step, macro recorder, and advanced scripting. Experiment was conducted in three phases to test the functionality of the developed macros in accomplishing all three tasks. The result showed that advanced scripting macro is the only type of macro that works. It is also the most suitable for the task because it is quick and easy to create and can be used to develop flexible and natural voice command. Since the output of macro is a series of keystrokes, which forms a syntax for the robot program, macros developed by the Dragon software can be used to communicate with virtually any robots by making an adjustment on the output keystroke.

Library of Congress Subject Headings

Human-robot interaction; Speech processing systems

Publication Date

5-13-2016

Document Type

Thesis

Student Type

Graduate

Degree Name

Manufacturing and Mechanical Systems Integration (MS)

Department, Program, or Center

Manufacturing and Mechanical Engineering Technology (CAST)

Advisor

S. Manian Ramkumar

Advisor/Committee Member

James Lee

Advisor/Committee Member

Alan Raisanen

Comments

Physical copy available from RIT's Wallace Library at TJ211.49 .K43 2016

Campus

RIT – Main Campus

Download

Share

COinS