Bill Daniels wrote:
I've been carefully watching the MEMS AHRS and IMU units for a while. The
makers claim good performance but charge high prices.

There is another way to achieve attitude sensing and that is multi-head GPS
receivers that measure carrier phase difference between the antennas. If
you combine MEMS AHRS with GPS attitude the two independent systems
complement each other such that the AHRS provides "coast-through" attitude
during GPS dropouts. It's not cheap but 0.1 degree resolution is possible.

There is no question that your idea will work, and work VERY VERY well.
NASA did some experiments and the results were WAY better than a
conventional AI. Just not quite ready for the consumer market yet -
price. But definitely has the potential to become a contender.

BTW, ref: the idea of putting multiple units in a "tell me thrice"
configuration. That would help, but except for the benefit of
redundancy, doesn't gain as much as one would think. They all tend to
suffer similar problems.

What we did do, which helped quite a bit, was characterize the units
over temperature. Although the variation between units is significant
(hence the benefits you noted), and the change with temperature is
DRASTIC, for any individual unit the change with temperature tends to
be quite repeatable. This allows you to temperature compensate the
unit with a response curve fit in software. BIG help.

Still leaves the inability to differentiate slow drift from a slight
deviation from level (or straight). But then a conventional AH
(electric or vacuum) has he same problem.