top of page

Speech Recognition for ROS Based Mobile Robots

 

Final Year Project for BEng Electronic Engineering

to About

    About Me    

My name is Feng Li and I have been studying Electronics Engineering in Middlesex University London for nearly two years. 

This website is for my Final Year Project about what I have done so far and what I need to do in the future.

Hope you guys can find something interesting here!

to Services

   About Project   

Planning 

 
First of all, I did some research about how to use ROS to achieve voice control on mobile robots and how to implement ROS voice recognition on Raspberry Pi3.
ROS is really convenient for beginners who don't know much about how to drive a normal robot around, so that's why I choose to apply it 😝
I found a ROS package which has voice recognition core function wrapped in, called PocketShpinx.  It is really powerful. People can even put whatever the words they want the machine to recognize using an online compiling tool to generate their own corpus.
Then I made some simulation tests using that package with Ubuntu virtual machine and a simulator called Gazebo and it worked perfectly.
After that, I built a dictionary with my own corpus using that online tool and the whole process was pretty quick. I installed ROS on Raspberry with PocketSphinx containing my dictionary as well. I finally made it work on TurtleBot which is a functional mobile robot based on ROS. 
Developing
Thinking
Since TurtleBot is a very functional robot with a high resolution rate camera, I want to achieve some more complex or interesting functions combined with voice control.
In the future, I also want to expand my dictionary with a big amount of corpus for more controlling choices.
to Work

    Works on It    

Network Setup

All Videos

All Videos

Middlesex University, The Burroughs, Hendon, London NW4 4BT

 

  • Facebook Social Icon
  • Twitter Social Icon
  • Google+ Social Icon

© 2023 by Name of Site. Proudly created with Wix.com

Success! Message received.

    Contact    

bottom of page