Skip navigation

Category Archives: Uncategorized




uniwheelisorender

Originally uploaded by Transforminglemons

This is a test post from flickr, a fancy photo sharing thing.

I have been working on the design of this project for close to two years, through many iterations and much trial and error, I am finally done with prototypes, testing, and concepts and ready to begin my final build.

Here is a rendering of my final design:NineFour Front

NineFour is a four legged, nine degree of freedom tele-presence quadruped.

The following is a breakdown of the NineFour features.

  • Qwerk  Robot controller
    • ARM9 RISC processor, 32 Mbytes SDRAM, 16Mbytes flash memory.
    • Linux 2.6
    • WiFi networking
    • Webcam support
    • 16 RC-servo controller
    • 8 12-bit analog inputs
    • 2 RS-232 ports
    • USB 2.0
    • Built-in audio amp for playing MP3 & WAV files
  • 9 Hitec HS-475 HB servos
  • 4 Hitec 645MG servos
  • 1/4″ laser cut polycarbonate
  • Pan/Tilt web cam
  • Foot contact sensors
  • 1 Ping Ultrasonic range finder
  • 4 Sharp IR distance sensors
  • Webcam & Laser for 3D object modeling on demand (using software from www.david-laserscanner.com)

The NineFour will be primarily controlled by a Ipad/Iphone app, but will be able to be controlled from anywhere on the globe with access to the internet as long as NineFour has access to a open WiFi network.

My ultimate goal is to create a website where anyone connect to and take control of the robot. A time limit would be imposed to allow many people the opportunity to control it.

In time I would like to add other bots to the system and allow for multiple users to control different robots.

The addition of 3D laser scanning has many possiblities as well. With the use of an onboard laser and webcam, it will be possible to sweep across an object and export a very accurate 3D model for manipulation and viewing. With the eventual addition of a 3D printer, these objects could be reproduced at will.

Sensor input comes from the 4 IR sensors mounted on the top of the robot. These sensors are mounted on a pivot and linked to a servo. These four sensors sweep 360 degrees, with the data being recorded every 3 degrees. This results in a radar-like display locating any obsticles in the surrounding 60 inches of the robot.

Radar

Navigation is achieved by one of two methods. First is manual control, essentially having the operator control the speed and path of the robot.

 The second method is autonomous navigation in which NineFour will navigate through a space, or play motion detecting security.

Below are some images of my progress, starting off with my first batch of laser cut parts:

Below  is an image of an early prototype rendering:

Here is an assembled prototype:

This prototype was designed as a proof of concept, and after initial testing, I determined the weak points in the design and corrected them in the lastest 3D model.

Some of these weak points include:

  • Too much flexing in the legs, solved by adding vertical supports.
  • Too much play in the leg linkage, solved by using gears instead of links.
  • Extreme battery drain, solved by adding higher power battery setup.

 Stay tuned for more updates!