Gyro and Accelerometer IMU for UAV... / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Gyro and Accelerometer IMU for UAV...

1 2 next

Move Thread

LAN_403

Patrick

Tue Feb 07 2012, 07:30PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

is it possible to use the 3 axis acceleromters to minimize the natural unwanted drift of a 3 axis gyro?

Im trying to do the physics math in my head, but it seems like i may need an optical flow lock to hold a hover indefinitely...

2Spoons

Tue Feb 07 2012, 10:31PM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

Yes you can correct to some extent, but only for pitch and roll gyros. A mate of mine did it for a 2 wheel balancing robot, but that was only one gryo and two acc. He used a Kalman filter to do the correction.

Carbon_Rod

Wed Feb 08 2012, 02:05AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

To a certain extent this is true, but practical use requires a 3-axis Magnetometer with the GPS elevation data.

A basic pack of all three is usually around $50 to $150 these days:

Most loss of precision starts by placing the sensor near metal parts during calibration, or regulated power supply problems.

Also note LSM303DLM are only good for Â±1.3 gauss, and the earth's field is around 0.25â€“0.65 gauss. The axis sensor range is only 12bits, and thus the data range will max out somewhere around 393 - 1025. Despite the resolution limits to find the tilt compensated direction, most students find it is still better than trying for real-time angular rate predictions.

Cheers,

ConKbot of Doom

Wed Feb 08 2012, 02:29AM

Registered Member #509 Joined: Sat Feb 10 2007, 07:02AM
Location:
Posts: 329

the previous (now discontinued) version of this IMU featured a kalman filter and FIR filter to combine all the data

they dont mention it for the UM6, but i would certainly hope they would still since they say it outputs roll pitch and yaw also. Perphaps contacting them and asking about it may be of some use. The magnetometer can also be used as an absolute reference for orientation provided strong magnetic fields arent nearby (I.E. dont park the IMU next to the motor, or the power lines supplying it)

Also, get 2 low range accels, mount them both in the x or y direction on long arms, facing the same direction (I.e. x+ excites both the same way) feed the output of each to a differential/instrumentation amp, (i.e. kill the common mode signal) and youre left with the difference, which will be acceleration they see due to yaw.

Carbon_Rod

Wed Feb 08 2012, 02:49AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

The units with built in filters are about $150 to $200 on the url I posted.

However, chip bundles are generally more versatile for filter designs.
=)

Patrick

Wed Feb 08 2012, 02:56AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

i was thinking of this from

spark fun, but my CG, CoR, CoM are located above two servo motors, would that be a problem for the magnetometer? would a thin steel sheet help?

Perhaps the magnetometer could be moved to the tail? (probly not )

I was thinking of using ultra sound, and IR range finding for elevation.

Carbon_Rod

Wed Feb 08 2012, 03:28AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

* Ultrasound will not be reliable in wind, and your money would be better spent on pizza.
* IR range finding works well indoors, but is somewhat dependent on the surfaces.

The HMC5883L is a good little chip (the 5884 is a little more pricey), and when set to Â± 0.88 gauss mode it adds even better resolution. Just keep in mind passing over cars may pin your sensor every so often, and the sample rate is a little lower.

"Any" magnetic metal object (like pins/plugs/wires/bolts) should be kept as far away from the sensors as possible. Use nylon mounting screws/ties, flexible multi-strand >copper< hookup wire, and a plastic project box. Note, some suppliers build metal objects near/into these sensors, and they should be avoided for long term usage issues.

Patrick

Wed Feb 08 2012, 03:44AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

ill put the gyros and acclers near the COG, but ill put the magnetometer in the tail, where its all fiberglass and carbon fiber. That should be 19.5 inches from the nearest metal and magnets.

Edit: does carbon fiber count as metal? its really conductive.

Carbon_Rod

Wed Feb 08 2012, 04:41AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

Most carbon fiber composites are not magnetic as the material is somewhat corrosive to untreated metal surfaces.

Placing the sensor cluster away from the center of mass is generally not a major issue, but they should be clustered around a single axis for better consistency.

Patrick

Wed Feb 08 2012, 06:08AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Carbon_Rod wrote ...

Most carbon fiber composites are not magnetic as the material is somewhat corrosive to untreated metal surfaces.

Placing the sensor cluster away from the center of mass is generally not a major issue, but they should be clustered around a single axis for better consistency.

ok ill embed carbon fiber in the fiber glass, and its just the mag sensor that will be in the tail...

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob