NavList:

   

Reply

George Huxtable  said:
> Mike Wescott said-
> >george@HUXTABLE.U-NET.COM said:
> >> But on a surface craft there are all sorts of random accelerations
> >> taking place, in unpredictable directions, of a magnitude comparable
> >> with that of gravity, depending on the sea state. These get hopelessly
> >> jumbled up with the acceleration due to gravity. Can a three-axis
> >> accelerometer somehow unscramble these accelerations?
> >
> >I suspect that a suitable low-pass filter could be used on the outputs
> >to get a reasonably accurate indication of the direction of the constant
> >1 G accelation.

> No, I think that would not work. Trouble is, you need a FAST
> response to the tilting of the deck to correct the fluxgate compass
> quickly for changes of tilt as the vessel gets chucked about. Otherwise,
> the indicated course would vary wildly in rough weather, even if the vessel
> continued to travel in a straight line. That would upset the use of this
> compass in a self-steering system, which is presumably its intended use.
> Mike Westcott's suggestion would provide a good AVERAGE course, but that
> would not be sufficient for a feedback steering application.
>
> The problem is that the frequency-spectrum of the changes of tilt is very
> similar to the frequency-spectrum of the accelerations as the vessel is
> pushed about, both being caused by the same waves. That is why it would be
> difficult, if not impossible, to separate them by any sort of filtering.

> All is not lost, though. As I pointed out in an earlier posting, "gyro"
> devices exist [...]

With a gyro mechanism, granted, it's a much simpler problem to solve. I would
still argue that it's feasible without. Mounting the 3-axis acceleration sensor
near the axes of roll, pitch and yaw (especially roll) would minimize the
acceleration affects of those motions leaving relatively short impulse
accelerations, or longer accelerations that are much smaller than 1G. Both
of which can be filtered.
--
        Mike Wescott
        Wescott_Mike@EMC.COM

   

Reply