Tonkin, Kyle (2013) Make Garfield (6 axis robot arm) Smart through the design and implementation of voice recognition and control. [USQ Project]
|
Text
Tonkin_2013.pdf Download (2MB) | Preview |
Abstract
The voice is a powerful tool used in everyday life. It expresses our feelings, emotions and is unique to every individual. Developing a system to decipher its functionality and deal with the complexities involved in its digitalisation, becomes quite the task. Once harnessed, it has the potential to become one of the most popular methods of control throughout the world.
Garfield (a 6 axis robotic arm) is in need of an extra method of control that can essentially make it ‘smarter’. In trend with current popularity, Voice Recognition has been selected as the method for control and the processes taken to implement such a system are explored in the following dissertation. Its outlines the design, development and implementation of a working voice recognition system used in conjunction with robotic arm software. In essence, this project explores the application of such a technology to the industrial robot industry and tests its performance as a whole.
A signal processing methodology was investigated, designed and developed in order to implement the voice as a control tool on the robotic system. This included the research and development of a voice detection algorithm based on the principles of voice recognition and required integration with robotic control software to execute appropriate movement.
Existing literature was explored in order to understand and apply the concepts of voice recognition and integrate it into a system responsive to user utterances. The basic outline of the processing techniques used followed those involved with Mel Frequency Cepstral Coefficients (Kumar & Rao 2011) and the recognition techniques involved with Euclidean distances (Muda, Begam & Elamvazuthi 2010). In applying these techniques within the Matlab platform, some Voice Processing tools (Brookes 1998) were sort out and used. A number of final design parameters were evaluated and tested in order provide recommendation to further system users and set objectives for future work.
The results of system performance testing discovered that jogging the system via the voice controlled method was viable with further improvement to the system response and quality of the control algorithm. Furthermore, possible solutions to the blemishes in the system are explored and avenues for further research and development stated.
|
Statistics for this ePrint Item |
Actions (login required)
 |
Archive Repository Staff Only |
