|
|
| |
From:
Hobart, Tasmania, Australia
What I do besides building with MINDSTORMS:
When I’m not at home with my wonderful family (including Mitch, 5, and Rohan, 3), I teach Mathematics, Information Systems and Robotics (of course!) at Claremont College, a terrific senior secondary school in Hobart’s northern suburbs.
I’m also heavily involved in running the Tasmanian state finals of the competition that feeds into RoboCup Junior Australia http://www.robocupjunior.org.au
Why I signed up for the MDP:
1. I thought it would be cool to be part of a community that was on the “inside” and have the opportunity to contribute ideas/suggestions at a time when they may be useful.
2. To play with, and develop teaching resources for, all the new capabilities of the NXT brick.
3. To explore the implications that the NXT platform will have for the RoboCup Junior Australia competition.
(I never really thought that I would have the remotest chance of being selected… woo hoo!)
Experience with MINDSTORMS:
I bought the RIS 1.0 at the end of 1999 as a reward to myself for completing my teacher training. I had lots of fun playing with it during the school holidays, and wanted to bring it into the classroom, but initially couldn’t see how to justify it educationally… duh! When the first RoboCup Junior Tasmania state finals were held in 2001, I realised that this was the opportunity that I had been waiting for. Since then, I’ve conducted a range of introductory robotics courses using MINDSTORMS and ROBOLAB for teachers and students from about 10- to 16-years old. [More recently, I’ve even introduced basic programming concepts to 5- and 6-year olds using the NXT-based robot scorpion, Spike!]
NXT Project: TwisterBot 
|
|
The spinning mechanism
|
Right or left hand (or foot?) blue!
|
|
|
Twister program
|
|
About this NXT Project:
Let’s be honest, I wasn’t selected for the MDP because of my skills in building, but hopefully this project will illustrate how an idea for a new robots can sometimes come from something very simple.
When my young boys and I play Twister, we have a problem... we all want to play! My solution was to create TwisterBot
to automatically spin the spinner and then call out the result.
For the building, I was surprised how easy it was to devise a structure that held the board so effectively. A touch sensor attached to the motor is used to detect when the arm has returned to ”12 o’clock”.
As for the program, I had originally intended that the spinner would rotate for a random period of time, and then the program would respond to the position. I soon realised, however, that it was much easier to ”cheat” and instead randomly select the outcome first, and then have the spinner go directly there. Once the spinner has arrived, the result is given (e.g. ”left foot, green”). We recorded audio samples for the directions, body parts, and colours (‘left’, ‘hand’, ‘red’, etc.) and used John Hansen’s Wav2Rso (http://bricxcc.sourceforge.net/utilities.html
) to convert them into *.rso files.
In between spins, my original plan was to use the sound sensor to have the TwisterBot
wait until we called out that we were ready. This was, of course, a bit of problem given the noise we were making getting into position. The next time I set this up I think I’ll use timer (perhaps with a decreasing time between spins as the game progresses) and/or a touch sensor to initiate the next spin. I wonder if a blue tooth device could help in some way...?
The LEGO Group and Milton Bradley will and cannot take any responsibility for the TwisterBot robot and its functionality in combination with the Twister® game.
NXT Project: Trent – a RoboCup Junior Australia Soccer Goalie 
|
|
Back View
|
Dribbler
|
|
|
Dribbler Front View
|
Side View
|
About this NXT Project:
Trent is a RoboCup Junior (RCJ) soccer goalie that uses stereo vision to track and chase a robo-soccer ball and then uses a magnetic compass and ultrasound to reposition itself in front of the goals.
My involvement with RoboCup Junior (http://www.robocupjunior.org.au) was no doubt one of the reasons that I was selected to be in the MDP, so it is no surprise that one of my first projects was to create a robot capable of playing RCJA Soccer.
Given how much bigger the NXT motors and sensors are compared to the RCX, I wanted to see how easy it would be to meet the current size restrictions using the NXT. The rules state that each robot, ”must fit inside an upright 22cm diameter cylinder and be no more than 22cm in height”.
Additionally, it seemed to me that the rubber wheels that came with the NXT would make an effective ‘dribbler’. That is, a ‘rotating drum that imparts dynamic back spin on the ball to keep the ball against the robot’ (RCJA Soccer Rules).
I found it a bit tricky to find an arrangement of the motors that worked, but eventually came up with something that was satisfactory. Once I had found this, however, I decided that a goalie didn’t really need a dribbler anyway, but instead should have stereo vision, and use the ultrasound sensor to find its position in front of the goals. I opted to keep the 3rd motor just in case I decide to turn Trent into a ‘striker’ at some point in the future.
I used ‘subsumption architecture’ as the basis for Trent’s programming. I believe this approach was developed by Rodney Brooks originally, but I came across it in Knudsen (1999), “The Unofficial Guide to LEGO MINDSTORM Robots”. I nearly had this completed using the LEGO MINDSTORMS NXT SOFTWARE, but found that it was easier to finish it off using RobotC (http://www-education.rec.ri.cmu.edu/robotc).
| |
|
|
|
|
|