Recommended Gyroscope for Orientation Measurement?

Is there a reason not to use MEMS accelerometers? You can use your 2 analog channels for the analog outs of an ADXL202 or similar.

I think a better bet would be to first improve the resolution of your encoders. A gyro is going to give you an angle rate which you will need to integrate to find the angle. Since gyros drift, you will soon have an angle error which will be much larger than your current encoder resolution. You will also need to figure out a way of combining the gyro data with the encoder data (which one will you believe more?) and this will lead you into Kalman filters which are interesting, but might not be the simple improvement which you have in mind.

If you really want a gyro, Tokin makes some including the discontinued CG-16D and Analog Devices makes some. There are also RC gyros which output pulses you can decode.

Mitch Berks> Hi all,

Gyroscopes detect rotational acceleration. Accelerometers detect linear acceleration. A complete inertial measurement device would make use of both.

dunno about recommend... I've not tried them but I've turned these up in the past.

regards, colin

If you haven't seen it already, a great paper on combining odometry and gyros (gyrodometry) is here:

Jeff Kroll

Actually, most sensors that are advertised as gyroscopes or gyro sensors detect rotational velocity, not acceleration.

It's true you need to combine them to make an accurate orientation sensor. Use the gyro mostly for fast motion (high-pass filter) because it'll drift when standing still, and use the accelerometer for the steady-state or slow-changing case (low-pass filter).

Randy

I've been trying to get a hold of some the Tokin gyro's, but they seem to be in short supply. I've been waiting for Bravo Electronics to get some since mid October. Anybody know where else they can be had?

I'd play with the Analog Devices gyro, but they only seem to come in BGA packaging which seems difficult to use from the hobbyist point of view. Any suggestions there?

Thanks,

Kevin

Analog Devices offers it on an evaluation board with non-BGA pins. They call them ADXRS150EB, ADXRS300EB and ADXRS401EB. You can also read about how someone handled mounting the BGA package:

The Tokin CG-16D is discontinued, but I'd be interested in what you find out about sources for current versions.

Mitch

Kev> I've been trying to get a hold of some the Tokin gyro's, but

STMicroelectronics used to sell gyros ("rotational accelerometers") for use in disk drives, washing machines, etc.

Not sure of the current status of their offerings.

I'm aware of this, but the OP is asking about "orientation measurement". The simplest way to determine the orientation of your robot is merely to measure the gravity vector relative to some reference surface of the robot. And an accelerometer is cheap and simple.

Correct. However, it should be noted that not all accelerometers are created equal. If you want to measure the gravity vector, you'll need a MEMS accelerometer, as piezo-based sensors can't detect static accelerations.

I would second Mitch's recommendation - increasing the encoder resolution will likely give you the greatest performance/cost benefit. To minimise your slippage problem, you could mount the encoders on separate wheels which are not driven by the motor - but are spring loaded to contact the ground. For a wheelchair type configuration these encoder wheels could be placed in line with the main drive wheel axis so they don't slip.

You might also consider an optical mouse sensor which can give quadrature output's for X and Y. Limitations are speed (40 ips) and i guess optics to deal with rougher terrain.

Kevin