MNR2.0

Okay, I took my working list of changes in MNR1 and applied almost all of them in MNR2.0. The only thing I did not fix was the orientation of the NXT Brick. I left it flat in MNR2.0. The tracked base was very successful in this version and I plan on diagramming the track layout in a LDraw and perhaps build up some instructions for building it again later when I need that type of base. I have had one problem with the tracked base since the very first use, it does not track straight. I'm going to try some program tweaks to try and improve it. In the current version it deviates 10-15 degrees in 5 feet.

I moved the third motor out of the drive cluster and made it drive a sensor platform. I used a very simple gear system to turn the ultrasonic sensor left and right 90 degrees after the sensor has stopped the robot's forward motion. As with most gear driven platforms I've tried there are problems with gear slippage and accuracy on the turns. I need to work on this somewhat.

Alright, with the changes from version 1 to 2.0 I think I have made good progress. I have a more stable system that rotates the sensors not the entire robot. The tracks are more stable but still do not travel straight. I did buy the Mayan Adventure book and I am reading the first part now. It uses the NXT-G system that comes with the NXT. I'm going to try resetting back to the original firmware and give that system another try. I like the Java kernel but I'm not sure the innacuracies in the tracks are not caused by the motor classes.

2.1 plans:

1. Target the sensor cluster and replace/simplify it.

2. Get the brick up out of horizontal position.

3. Try the NXT-G IDE.

Maze Navigating Robot 1

I've put the card dealing robot on the back burner for a while. I am trying to make a robot that can navigate a maze. I'm going to research and try sensors from both LEGO and other manufacturers to make a robust robot. This line of robots will be the MNR series. I started with MNR 1.

As you can see, I am using a tracked design instead of wheels in either the 4 wheel, 6 wheel or 3 wheel designs. I might revisit this in the future, but I made this decision because the three wheel designs were very unstable and I wanted stability. With MNR1 I had problems with the tracks pulling to the left or the right. I belive it has something to do with the wheel base calculation required for the combine motor control in Lejos. The pilot object will drive the two tracks as a single unit handling the turns and the forward/backward for me, but I have to get the calculations right. In addition I had problems with the tracks bending away from the frame because of track tension.

I used a touch sensor with a probe on the end to determine when the robot should stop and I used the light sensor to check intensity to the left and right on a stop . These sensors were mounted on the center line of the body. I had problems with this design because the touch sensor was too high and missed low objects. I also disliked the way the whole robot had to turn to determine which way to go next. I will use the ultrasonic sensor next time.

Overall this was an interesting starting point. I learned a lot from this version.
1. I want a better tracked platform. I want to focus on making the motor and tracks configuration better.
2. I want to seperate the sensor bed motor from the cluster of motors.
3. I want to try the ultrasonic sensor for stop position control.
4. I want the sensors to rotate not the robot.
5. The NXT Brick needs to be angled better to improve connection to the PC.

I'm going to review more sensor options and look at the book out by APress on navigating mazes using the NXT. I believe it is called The Myan Adventure. It looks like a good resource.