(The) quadruped robot TITAN-VIII control by gesture and speech recognition = 제스처 및 음성인식에 의한 4족 보행 로봇제어

다국어 초록 (Multilingual Abstract)

Abstract
Nowadays, robots are so important in many fields of our life such as
military, industrial products counter, explore, sport, traffic monitoring. Most
the designed and developed of robots are to work in extreme environments
such as ocean bottom, space and hazard prevention. It is not easy for the
people living there because of the high atmosphere, the low temperature and
the high risks for the life of people. However, the robots are capability of
replacing people to reduce the risks.
The robots are also separated into two types of Wheeled-robot and
Legged-robot that depends on the mobility of the robot that is a wheel or a
crawler, respectively. Legged robots (or Walking robots) have a potential to
pass over certain types of terrain in a more efficient and stable manner more
than the others, using wheels or tracks. Other advantageous properties of
walking robots are that they cause less damage to the terrain, and they are
omnidirectional, which gives them an advantage in moving through cluttered
and tight environments. Besides them, the great development of computer
vision technique can be applied into robots. The computer vision includes
methods for acquiring, processing, analyzing, understanding the digital
image and extracting information from images to control a real system of
user. In the automation system, the image information can be used as sensor
data to detect parameters of target image and control object tracking,
following system. The combination between robot and computer vision
technique is a today’s trend to make a smart system.
A description of the implementation and experimental results are
provided using the quadruped robot TITAN-VIII.
The main contribution of the dissertation is in the analysis of the
stability of balanced quadrupedal gaits statically and how statically balanced
walking and the various posture of the robot is discussed to keep the robot
always in the statically stable without mathematical model. Because of
without a complex mathematical model of TITAN-VIII, a controller is added
into system to help the robot moves to target smoothly and exactly.
Based on above works, the math formulas are presented to determine
the joint angles of the robot and simulate on Matlab & Simulink software.
After that a program is created to recognize human gesture and speech.
TITAN-VIII is equipped with one Microsoft's Kinect sensor which includes
3D scanner, color camera and depth sensor. TITAN-VIII is tested in real
environment with normal light condition. Results of this dissertation are
extremely impressive when TITAN-VIII can automatically and quickly
avoid obstacles. From the success of experimental results, this control
method can be applied into other similar smart robot easily and widely to
develop new useful systems.

목차 (Table of Contents)

• Table of Contents
• List of tables
• List of figures
• List of photos
• Nomenclature
• Abstract
• Chapter 1 Introduction
• 1.1 Background of the research ................................................................... 1
• 1.1.1. The importance of robot in human life ......................................... 1
• 1.1.2 Why study Gesture recognition and quadruped robot ................... 5
• 1.2 Scope of research ................................................................................... 6
• 1.3 Target and method of research ............................................................... 7
• Chapter 2 Construction of TITAN-VIII
• 2.1. Introduction to TITAN-VII ................................................................... 9
• 2.1.1. The feature of TITAN-VIII ........................................................... 9
• 2.2.2 The main part of control system .................................................. 10
• 2.2 Mechanical structure of TITAN-VIII .................................................. 23
• 2.2.1 Leg mechanism ............................................................................ 24
• 2.2.2 The gait and stability of the quadruped robot .............................. 26
• Chapter 3 Simulation of sequence of swing leg
• 3.1. The sequence of TITAN-VIII legs ...................................................... 30
• 3.1.1. The importance of generating a gait for TITAN-VIII ................ 30
• 3.1.2 The posture in the straight-going moving ................................... 32
• 3.1.3 The posture in the turning gait moving ........................................ 37
• 3.2 Control of the joint positions and control algorithm ............................ 39
• 3.2.1 Control of the joint positions ...................................................... 39
• 3.2.2 The control algorithm .................................................................. 42
• 3.3 Generation sequence of legs simulation .............................................. 44
• 3.3.1 Designing TITAN-VIII model on SolidWorks ............................ 44
• 3.3.2 Position PID controller of revolute joints .................................... 52
• 3.3.3 Experimental simulations ............................................................ 60
• Chapter 4 Gesture recognition using Kinect camera for TITAN-VIII
• 4.1. Introduction to gesture recognition .................................................... 67
• 4.2 Architecture of the system ................................................................... 68
• 4.2.1. The software ............................................................................... 68
• 4.2.2. Kinect camera ............................................................................. 69
• 4.3 The principal of human Recognition system ....................................... 72
• 4.3.1 Gesture recognition base on the human skeleton tracking .......... 72
• 4.3.2 Skeleton tracking method ............................................................ 74
• 4.4 Implementing Human Recognition on Visual Studio .......................... 78
• 4.4.1 Setting up a Kinect SDK .............................................................. 78
• 4.4.2 Implementing Human Recognition .............................................. 79
• 4.4.3. Implementing speech recognition ............................................... 89
• 4.5 Experimental results of gesture recognition program .......................... 95
• 4.5.1. The control program of the robot ................................................ 95
• 4.5.2 Setting up robot’s hardware ......................................................... 96
• 4.5.3 Experimental results .................................................................... 99
• Chapter 5 Conclusions
• 5.1 Conclusions ........................................................................................ 108
• 5.2 Future research ................................................................................... 109
• References .................................................................................................. 111
• Appendix .................................................................................................... 115
• Acknowledgement ..................................................................................... 131