Let's Make Robots!

I love the Robotics since I was a young boy: "Mom, when I'm BIG I will create a robot that washes the floor, washes the clothes and irons them for you"

Now I'm "BIG"! I studied Computer Engineering with a specialization in Robotics, I'm a Computer Vision expert and I know a lot about Artificial Intelligence and Machine Learning, so it is time to complete what I called... 

... the "MyzharBot Project"


MyzharBot will be your personal robot. You will be able to speak to it and it will recognize you and it will reply to your vocal commands.

MyzharBot will be a full autonomous robot, but you will be able to control/monitor it by your smartphone/tablet (Android, iOS and Windows) via WiFi, bluetooth and remotely by Internet.

Main Features

  • BrainPandaboard ES, dualcore ARM Cortex-9 miniPC
  • Operating System: Embedded Linux Ubuntu based
  • Power: LiIon 5000mA Turnigy Battery
  • Motors: 2x powerful Micro Motors series e192 – model e192-12-18 - Link - Datasheet
  • Motor Encoders: 2x CUI INC AMT102-V - Link - Datasheet
  • Motor Logic Unit: Robocontroller V2 - Link
  • Motor Power Unit: Roboponte-MD V2 - Link
  • Tracks: 2x Tracks LynxMotion 3″ width - Link
  • Chassis: Laser cut plastic chassis. Project on Ponoko site

Project State

MyzharBot's prototype is actually under developing. You can see a few of videos made during preliminar tests on Myzhar's channel on YouTube.

The chassis is defined and a first project is ready, but a few modifications must be studied to let it be more robust and let the hardware be well suited.

A lot of work must be done on the software side. Linux system must be configurated and MyzharBot software must evolve such to include Artificial Intelligence. The software framework will be based on ROS, a Robotic Framework developed by WillowGarage.

Computer Vision processing will be based on OpenCV library, the most complete and performat Computer Vision Library available. I love it!

The Control Unit (GUI) is ready for Desktop PC and must be ported to mobile platforms.

A draft scheme of the architecture:

One encoder suited on the motor shaft:

The brain, Pandaboard ES:

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Myzhar's picture

OpenCV is the state of the art of the Computer Vision. It is free and it is used very hard in research and it is beginning to enter in the industrial world. I use OpenCV since 2003 when I started to work to my Engineering thesis project. It was at the alpha stage and it was completely written in C.

Now OpenCV has been "translated" in many languages and finally hat got C++ interface that improved a lot its utilization.

A lot of great industrial houses collaborates to OpenCV development: Intel, NVidia (OpenCV supports CUDA parallelization on GPU), Google, Toyota and more.

mogul's picture

Looks pretty advanced. I have always wanted to play with computer vision in some way, and being a linux user OpenCV seems to to be an obvious choice. Unfortunately I'm nowhere a computer scientist, so the learning process is a quite steep uphill journey.

 

Myzhar's picture

Using OpenCV is easiest than it seems. If you have basis of programming you can write Computer Vision algorithms in a little time.

OpenCV in its latest versions has a Python interface too that let you write code with a minimum effort if you do not know C or C++..

On my site there are a few tutorial about OpenCV and I'm publishing some basic video tutorials (in italian) on another site: www.opencv.it.

6677's picture

Looking on google there seem to be OpenCV bindings/wrappers for a huge variety of languages.

Seems a good choice on any platform.