| User | Post |
|
12:18 am
July 29, 2008
|
PirCapAdi
| | | |
| Member | posts 24 |
|
|
Hi all,
A quick update. I finally got time to finish the odometry and basic way point system. It still has a pretty big error, but some more tinkering should fix that.
http://www.youtube.com/watch?v=zLDZUU-Bh-Q
This is a video of it navigating a set of points in my apartment, no other sensors, only quadrature encoders on the wheels.
Enjoy.
Adi
|
|
|
6:51 am
July 29, 2008
|
ash
| | | |
| Admin
| posts 67 |
|
|
Cool Adi, that base is a lot smoother than I was thinking it was going to be. I have been meaning to buy one for ages but then kept thinking I should make one instead (which I didn't actually do  . Might have to look into it again.
Good work!
|
|
|
6:59 pm
July 29, 2008
|
robotman
| | | |
| Member | posts 16 |
|
|
Excellent!
How many points did you use? Did you just create a set point when the base needed to turn? And it nearly took out the wall on that first corner too!
|
|
|
10:01 pm
July 29, 2008
|
PirCapAdi
| | | |
| Member | posts 24 |
|
|
Should have been three points, but forgot to remove one, thus the circle at the end. It did hit the wall several times in test runs, I did that run about 5 times before I was confident enough to stand back and let it do it on its own.
The robot uses a very simple state based system, with way point “tracking” as a state. While in this state it will move to a set way point, when it reaches the way point it signals that it has and the state system then passes the next way point if there are any, else is changes state to “not tracking”. Thus simply creating way points where I needed it to turn worked great.
I am now working on integrating a CMPS03 compass which should reduce some of the error caused by the skid steer.
This system, odometry with compass and IMU, will form the core of it's navigation system, GPS will only be used to correct errors. Relying on GPS seems to be the biggest downfall of most robots entering Magellan events.
|
|
|
8:32 pm
October 8, 2008
|
PirCapAdi
| | | |
| Member | posts 24 |
|
|
Been a while since there has been any robotics talk so thought I would post a quick update…
I dropped the ATMega32 control board, now using a WiringIO board with custom perfboard IO breakout header (WiringIO "Shield"); will do a proper PCB once it is all proven.
With the additional IO on the ATMega128 I have been able to drop the 4 TinyQED’s and the I2C servo controller. I have also replaced the 4 HB25’s with two Sabertooth dual 5A controllers. The I2C LCD has also been switched for a traditional 4bit LCD. I have got an Ocean Server tilt compensated compass and an ENC28J60-H SPI Ethernet interface; I love Sparkfun.
I have also prototyped a new 5V switch mode PSU.  .jpg
All small changes, but all at once has put me back about 2 months. Initial tests shows that the system is substantially quicker and more accurate. Just need to transfer the last few connections from the breadboard to the shield, will post some pics and videos once that is done.
Adi
|
|
|
2:49 am
October 9, 2008
|
ash
| | | |
| Admin
| posts 67 |
|
|
Good work mate Good to see some non web-development stuff on here again hehe.
Man, everytime Sparkfun update their page it's like christmas! Their stuff just keeps getting better and better.
The new helical GPS I saw on there looked good. 4 Hz operation and from the description they seemed to imply that it will work well in a lot of places others wont be able to get a lock (like under a car seat even). And the wifi module, awesome! Must. not. buy. everything….
Keep up the good work!
|
|
Search 
The Forums
No Status Selected
Print this Post 
