Tuesday, 10 March 2009

chess bot

This task is a semi lead on from our previous task we had to apply our chess piece protocol to a Lego NXT kit, when I first heard this I was excited, yay we get to program our brick to travel across a giant chess board with our protocol I thought out in my head. but then the horror hit we had to set it up to navigate to the year two studio. a path which involves a right corner then a left corner with a wall at a 45° angle. previously on BCT my group decided that we would create a piece that moved forward 2 then diagonal 1 to its left

The group originally went in a direction of creating a unique model instead of the default working base of the tribot, our first design was a stretch frame with two leading wheels the brick slung in the middle with a supporting wheel. the fault of due the fragility of the frame the chassis would tort making the support wheel unbalanced making the robot curve left on the straights.
after trying a varied array of improved rudder wheels and sleds we then decreased the wheel base, trying to shift the weight balance onto the lead wheels instead of the rear support wheel. resulting in a design dubbed the trolley. which then had claws added to make it less docile.

We started off with a sonar sensor, but fairly late in testing we discovered why it seemed to miss the evasive statements in the program loop at around angles greater than approximately 45° to the left the value for distance would show a reading of 0 when the angle was to 275 ° the wall it would jump from 215° and 0°. due to this when the robot hits the wall at angle it would just dry hump the wall or try and climb it. due to only having one touch sensor we decided to mount it on the left had side since the whole obsession with turning left.

Another one of the issues we struck was trying to tune the angle at which the unit turned which was more of a trial and error test although getting it to turn on the spot was elusive to this time, although I do have a code sequence to potentially alleviate this.

We were the first group to head out into the hallway, by coming up with a flawed robot it allowed us to improve, this assignment wasn't so much a means to succeeding the easiest way possible. (i.e. coding in the instructions, you know which groups your are) but to learn the ropes and practice solving issues that arise from making the NXT autonomous, much like coding its a series of solutions to solve a problem which then create their own problems in a repetitive cycle till you can pass said problems of as a feature.

Labels: ,

2 Comments:

Blogger evren said...

Ryan, very well done with documenting and describing the processes. Keep up the good work in blogging!
Evren Uzer

11 March 2009 at 4:40 pm  
Blogger evren said...

And also double thanks to you for following others' works and commenting on them in a very contributing way!
Evren

11 March 2009 at 5:09 pm  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home