Comment

Comment by Niall Dempsey on July 20, 2012 at 1:04am

Just placed an order for APM2, uBlox GPS and 3DR Radio Kit

Comment by Niall Dempsey on July 18, 2012 at 12:52am

I'm in a different time zone so I'm a bit out of sync with the discussion.

I'm currently using  a HMC6352 compass and had to move it 2 feet away from the motors, even at 8 inches away from the 2 * 11 amp motors it was useless. As I said previously the GPS navigation still has some glitches and I suspect this is due to the compass. Does the ArduRover code make use of the Gyro in its calculations?

Comment by KM6VV on July 17, 2012 at 2:39pm

@Michael,

Thanks for the "Nickle Tour"!  I had read about your "Digital Bus" before, but not this blog.  Looks like a lot of information to digest!  I'm looking forward to studying it.

 It'll be great to have you chime in on questions.

Thanks!

73's, Alan KM6VV

P.S. How do we put Ham Radio on these 'Bots?

Comment by Bot Thoughts on July 17, 2012 at 1:56pm

@Alan PS: I would be happy to go into more detail here about anything in particular to whatever degree of depth time and interest permits :)

Side note: my call sign is K0FSJ

Comment by Bot Thoughts on July 17, 2012 at 1:53pm

@KM6VV - Thanks! I was actually able to post the 2nd fastest raw (real) time. Alas if only I'd gotten that 30 second hoop bonus heheh. :)

In fact, I didn't use APM at all.

Instead Data Bus runs custom hardware and software. I rolled my own navigation system using Kalman Filter for heading estimation. Here's more details on the bus: http://www.bot-thoughts.com/2011/05/data-bus-nickle-tour.html which includes links to details on the various tests. calculations, and simulationsI did, as well as more detailed info on the heading / position estimation, and a ton of other stuff.

Comment by KM6VV on July 17, 2012 at 1:35pm

@Michael,

I did read that you did quite well at AVC with your Bus!   Was that all APM2, or did you use a supervisory uP to do the obstacle avoidance?  We'd sure like to hear more!

Alan

Comment by KM6VV on July 17, 2012 at 1:32pm

Hi Michael,

Sounds like you've done a lot of work on your 'bot and navigation! 

Yes, I was just reading up on the 6000. Having motion fusion in this chip should simplify things.  Does the APM2 interface magnetometer info into the 6000?  I didn't notice it on the schematic.  Although I still trying to get familiar the the APM2.  'Hope to get it on my new 'bot (still under construction). 

Your comments on the magnetometer reflect what I've been suspecting; I'd like to get mine up and away from the ground (as well as motors) too.

That turntable sounds interesting,but I won't have too much time to pursue the effects on the GPS  and magnetometers (I get sidetracked too easily).

Thanks for the comments!

Alan

Comment by Bot Thoughts on July 17, 2012 at 12:58pm

@ KM6VV: Oh, I'm Michael (bot thoughts is actually my blog). Truth be told, my robot hit 20mph top speed in the AVC. Turns at about 10mph or so (could probably do more). The only reason I didn't go faster is that I wanted to quit while I was ahead two nights before the event. :D  It wasn't at full throttle or anything. :)

The 3d corrections... well, with ArduPilot, it's all fused together to get attitude and heading using the DCM based algorithm. Accelerometer provides a down reference, magnetometer provides a north reference, the references unbias the gyros which are used as the primary means of estimating attitude. At least that's my understanding having gone through the code when I ported it to the mbed for my robot early on. So under sustained centripetal acceleration, without any compensation, it could throw off the heading estimate. I decided after doing some calculations to ditch 3d for 2d.

I used a turntable machine I devised with an encoder to calibrate both gyro and magnetometer, accurate to 2° so I could get a sense of how well my calibration worked. It'd be nice to have a commercial rig to test.

So I went 2d. I don't have a magnetometer but will add one if I can figure out how to deal with the errors I was seeing.

I may try raising my magnetometer (except actually I have a Sparkfun MinIMU-9 so have to instead use a separate compass and gyro module). 

Comment by KM6VV on July 17, 2012 at 12:16pm

@bot (name?),

Yeah, my spelling checker just doesn't know what I'm thinking either!

OK, maybe not so high a pole If the GPS is as hot as I'm hearing!  And it would be simpler if I don't have to cable up the magnetometer as well as the GPS.  GPS is already a separate module, so not much effort there.  And I want to be able to quickly detach it (or I could buy another one) in order to be able to "walk the field" before the competition starts (can't take the whole 'bot around), and get my waypoints with my own GPS.

Doesn't sound like you have much magnetic field to worry about.  We need either a commercial magnetometer, or an Arduino program to work as a meter with our sensor.  That way we'd know what we had.  I'm using two big brushed motors from SuperDroids (42s).  Electrical noise is probably more of a problem. 

The 3D corrections would be more for the compass, I think. 

Need to run inertial navigation to back up the GPS for when it has lost its signal.  Maybe that's not so much of a problem anymore! 

I'm thinking of making up a 12" x 12" x 20" Plexiglass box for my electronics, that would get the sensors up a bit above the motor deck.

How did magnetometer readings change over ground with the mag up higher?  Looks like I've got some testing to do. 

Wow!  >15MPH.  I think I set up my 'bot for about 6+ MPH at max motor speed.  I can change the ratio by selecting different timing belt pulleys.  

Alan

Yeah, APM2 is actually cheaper then APM1 now!  I almost bought a used APM1. 

Comment by Bot Thoughts on July 17, 2012 at 11:37am

inertial not internal. heh, typo. :)

• ‹ Previous
• 1
• …
• 55
• 56
• 57
• 58
• 59
• …
• 64
• Next ›
• Page