EDAboard.com | EDAboard.de | EDAboard.co.uk | WTWH Media

Using many cheap accelerometers to reduce error

Ask a question - edaboard.com

elektroda.net NewsGroups Forum Index - Electronics Design - Using many cheap accelerometers to reduce error

Goto page Previous  1, 2, 3, 4  Next


Guest

Sun Feb 10, 2019 8:45 am   



On Sunday, February 10, 2019 at 1:15:10 AM UTC-5, DecadentLinux...@decadence.org wrote:
Quote:
JS <js5071921_at_gmail.com> wrote in
news:adc045fe-69d3-4561-976a-f750a84b5256_at_googlegroups.com:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser ring
gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done
in a chip fab, it should be possible.

What makes you think they need to be 'together'? Simply on the
same superstructure should be enough.

Is it possible to make a commercial accelerometer with no export
restrictions by using such an array?

Is it possible to feesibly make such an array, cost, logistics,
etc. at all?

Or will ITAR or the like be
slapped on such a device once its accuracy is published in a
brochure?

ITAR restricts YOU the maker long before you go publishing product
performance numbers. If you make certain items, you 'learn' the
restrictions (long) before you get past single proto stage.


Perhaps you are saying this in a tongue in cheek way, but ITAR is one of those things where you are expected to know your device is restricted without anyone in the government telling you. Not only are you required to not sell it to the wrong people, but you are supposed to know that the people you sell it to won't sell it to the wrong people.

Rick C.

Martin Brown
Guest

Sun Feb 10, 2019 10:45 am   



On 10/02/2019 05:58, JS wrote:
Quote:
On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST), wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser ring
gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.


It might be interesting to try out say 16 to see how much of an
improvement is obtained and how much of it is systematic error.

Quote:
It sounds like a lot of components to solder together but if done in
a chip fab, it should be possible.


A 10 fold performance improvement with 100 or 128 might be worthwhile if
they are cheap enough to fabricate in bulk.


--
Regards,
Martin Brown


Guest

Sun Feb 10, 2019 2:45 pm   



On Sunday, February 10, 2019 at 4:16:24 AM UTC-5, Martin Brown wrote:
Quote:
On 10/02/2019 05:58, JS wrote:
On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST), wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser ring
gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It might be interesting to try out say 16 to see how much of an
improvement is obtained and how much of it is systematic error.

It sounds like a lot of components to solder together but if done in
a chip fab, it should be possible.

A 10 fold performance improvement with 100 or 128 might be worthwhile if
they are cheap enough to fabricate in bulk.


The problem is the square root relationship. That means you need N^2 more units to get an N factor improvement which means a *lot* more units.

Rick C.

Clive Arthur
Guest

Sun Feb 10, 2019 3:45 pm   



On 10/02/2019 09:16, Martin Brown wrote:
Quote:
On 10/02/2019 05:58, JS wrote:
On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST), wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser ring
gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It might be interesting to try out say 16 to see how much of an
improvement is obtained and how much of it is systematic error.


A UK seismometer company did (does?) just this, 16 MEMS accelerometers
per axis. I don't know what the improvement is, but I'm certain they'd
lie about it knowing the customer couldn't check. [That's why I only
stayed there a few months, shocking quality and shocking performance
exaggeration.]

Quote:
It sounds like a lot of components to solder together but if done in
a chip fab,  it should be possible.

A 10 fold performance improvement with 100 or 128 might be worthwhile if
they are cheap enough to fabricate in bulk.


Cheers
--
Clive

John Larkin
Guest

Sun Feb 10, 2019 5:45 pm   



On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS <js5071921_at_gmail.com>
wrote:

Quote:
On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to 1/sqrt(sample size), does having many accelerometers and then averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap accelerometers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly better
than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro (0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25 deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export restrictions by using such an array? Or will ITAR or the like be slapped on such a device once its accuracy is published in a brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8


If the MEMS parts use vibrating cantilevers, they would want to sync
up. I don't know if that is good or bad.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics


Guest

Sun Feb 10, 2019 5:45 pm   



On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin wrote:
Quote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS <js5071921_at_gmail.com
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to 1/sqrt(sample size), does having many accelerometers and then averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap accelerometers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly better
than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro (0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25 deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export restrictions by using such an array? Or will ITAR or the like be slapped on such a device once its accuracy is published in a brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to sync
up. I don't know if that is good or bad.


They might "want" to sync up, but I'm not sure they would. If the platform has rotational acceleration there would be a difference in the acceleration on each device depending on it's distance from the center. That would keep them out of sync.

Rick C.


Guest

Sun Feb 10, 2019 6:45 pm   



gnuarm.deletethisbit_at_gmail.com wrote in
news:12d7dd9c-e906-48a6-9c63-68ec33e58be5_at_googlegroups.com:

Quote:
On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin
wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS <js5071921_at_gmail.com
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser
ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if
done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export
restrictions by using such an array? Or will ITAR or the like be
slapped on such a device once its accuracy is published in a
brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial
Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008]
Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to
sync up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If the
platform has rotational acceleration there would be a difference
in the acceleration on each device depending on it's distance from
the center. That would keep them out of sync.

Rick C.


If the 'platform' it was mounted to was 'known' (by the software),
the positions and differences from the center reference of any arc
like motion of the platform would be part of it.

Does anyone remember the guy whom all on his own developed little
quad rotor 'birds' that would all act together in a swarm fashion?
He showed it off in an ice rink or such and they were extremely
controllable and all knew exactly where they were in reference to
each other.

Pretty sure DARPA snatched him up. They wanted his software and
the algos he used.

JS
Guest

Sun Feb 10, 2019 9:45 pm   



On Sunday, February 10, 2019 at 6:55:35 PM UTC+2, DecadentLinux...@decadence.org wrote:
Quote:
gnuarm.deletethisbit_at_gmail.com wrote in
news:12d7dd9c-e906-48a6-9c63-68ec33e58be5_at_googlegroups.com:

On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin
wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser
ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if
done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export
restrictions by using such an array? Or will ITAR or the like be
slapped on such a device once its accuracy is published in a
brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial
Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008]
Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to
sync up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If the
platform has rotational acceleration there would be a difference
in the acceleration on each device depending on it's distance from
the center. That would keep them out of sync.

Rick C.


If the 'platform' it was mounted to was 'known' (by the software),
the positions and differences from the center reference of any arc
like motion of the platform would be part of it.

Does anyone remember the guy whom all on his own developed little
quad rotor 'birds' that would all act together in a swarm fashion?
He showed it off in an ice rink or such and they were extremely
controllable and all knew exactly where they were in reference to
each other.

Pretty sure DARPA snatched him up. They wanted his software and
the algos he used.


It is not just the bulk motion of the body which you have to take into account, but also all the structural vibrations and their harmonics if you want to do averaging of accelerometer values over a distributed space.

For example, if you have three arrays of accelerometers on the nose, wing root and tail, all three will experience and report different accelerations if the drone is buffeted by wind and its wings experience a momentary boost in lift. Thing is you don't even know what the wind pressure-space-time distribution is exactly so you can't account for it.

It would take a supercomputing cluster to do a real-time dynamic analysis of all the accelerations from bulk + vibrational motion in your platform, even IF you knew to a high degree of accuracy what all the forces acting on your platform are.

The very act of trying to correct for systematic errors in a distributed accelerometer array introduces new errors in terms of the accuracy of your mathematical models.

But your idea of a spatially distributed sensor network for navigation sounds interesting. Migratory birds for example are fascinating because of their ability to fly intercontinentally and land in their previous destination last year, with great accuracy and no trial-and-error. I wonder if one bird would get lost whereas as the flock size increases, their likelihood of getting lost decreases?


Guest

Sun Feb 10, 2019 9:45 pm   



On Sunday, February 10, 2019 at 6:28:43 PM UTC+2, gnuarm.del...@gmail.com wrote:
Quote:
On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to 1/sqrt(sample size), does having many accelerometers and then averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap accelerometers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly better
than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro (0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25 deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export restrictions by using such an array? Or will ITAR or the like be slapped on such a device once its accuracy is published in a brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to sync
up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If the platform has rotational acceleration there would be a difference in the acceleration on each device depending on it's distance from the center. That would keep them out of sync.

Rick C.


If the body is rigid then s = r*theta, and you can take the time derivative twice for velocity and then acceleration.

Problem is that real life is not so easy. A long slender missile body for example would have very different transient accelerations at the tip and the tail. After you double integrate the accelerations to get displacements, over a distance of say 10 - 1000km range, your CEP becomes wide as the side of a barn.

The best thing really for such applications is to either locate all your cheap accelerometers into a small, rigid box, or use a more niche but higher accuracy accelerometer like a RLG.


Guest

Sun Feb 10, 2019 11:45 pm   



On Sunday, 10 February 2019 05:08:13 UTC+1, DecadentLinux...@decadence.org wrote:
Quote:
Sylvia Else <sylvia_at_email.invalid> wrote in news:gc9co2Fmqn1U1
@mid.individual.net:

Any approach using large numbers of less accurate parts is not
guaranteed to give you the accuracy you want.


One exception might be when paralelling resistors. 1% resistors in
paralell will generally be more accurate than the original spec. Maybe
due to the way precision classed resistor sets get matched and culled.
One can generally count on the members of the set to actually be more
accurate than the spec they claim to be at least as good as.


That is not my experience. Resistors are tuned in the process, which means that one lot will typically have more or less the same distribution, but it is offset from the nominal. If they measure its's within the specs (1%), then they do not alter the process to pull it in. They just press the big "GO" button

Thus more resistors in parallel gets you nowhere

Cheers

Klaus

John S
Guest

Sun Feb 10, 2019 11:45 pm   



On 2/10/2019 12:35 AM, JS wrote:
Quote:
On Sunday, February 10, 2019 at 8:15:10 AM UTC+2, DecadentLinux...@decadence.org wrote:
JS wrote in
news:adc045fe-69d3-4561-976a-f750a84b5256_at_googlegroups.com:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(sample size), does having many accelerometers and then
averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
accelerometers rather than one expensive one like a laser ring
gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro
(0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done
in a chip fab, it should be possible.

What makes you think they need to be 'together'? Simply on the
same superstructure should be enough.

Incorrect. On the same superstructure there will be additional structural vibrations whose amplitude and frequency are position-dependent. While these vibrations do not result in a net spatial or angular translation (you would hope so!) for the math to work the accelerometers have to be at the same location, the closer the better, sampling the same thing.


Which is why he is known as AlwaysWrong.


Guest

Mon Feb 11, 2019 1:45 am   



On Monday, February 11, 2019 at 9:43:30 AM UTC+11, John S wrote:
Quote:
On 2/10/2019 12:35 AM, JS wrote:
On Sunday, February 10, 2019 at 8:15:10 AM UTC+2, DecadentLinux...@decadence.org wrote:
JS wrote in
news:adc045fe-69d3-4561-976a-f750a84b5256_at_googlegroups.com:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

What makes you think they need to be 'together'? Simply on the
same superstructure should be enough.

Incorrect. On the same superstructure there will be additional structural vibrations whose amplitude and frequency are position-dependent. While these vibrations do not result in a net spatial or angular translation (you would hope so!) for the math to work the accelerometers have to be at the same location, the closer the better, sampling the same thing.

Which is why he is known as AlwaysWrong.


Not true. DecadentLinuxUserNumeroUno is called AlwaysWrong by krw because he's one of the many people who disagree with krw.

DLUNU can be irritating, and the name AlwaysWrong gets thrown at him from time to time by other people when he's gotten more irritating than usual - I've done it - but DLUNU is right some of the time, and more frequently right than krw.

He's done some interesting stuff, but he's almost as likely to fly off the handle as Phil Allison.

--
Bill Sloman, sydney


Guest

Mon Feb 11, 2019 2:45 am   



On Sun, 10 Feb 2019 08:28:39 -0800 (PST),
gnuarm.deletethisbit_at_gmail.com wrote:

Quote:
On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS <js5071921_at_gmail.com
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to 1/sqrt(sample size), does having many accelerometers and then averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap accelerometers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly better
than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro (0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25 deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export restrictions by using such an array? Or will ITAR or the like be slapped on such a device once its accuracy is published in a brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to sync
up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If the platform has rotational acceleration there would be a difference in the acceleration on each device depending on it's distance from the center. That would keep them out of sync.


Would it? If they're on the same platform, the coupling is the same
no matter what additional force is on them (superposition).


Guest

Mon Feb 11, 2019 3:45 am   



On Monday, February 11, 2019 at 11:51:42 AM UTC+11, k...@notreal.com wrote:
Quote:
On Sun, 10 Feb 2019 08:28:39 -0800 (PST),
gnuarm.deletethisbit_at_gmail.com wrote:

On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS <js5071921_at_gmail.com
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to 1/sqrt(sample size), does having many accelerometers and then averaging their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap accelerometers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly better
than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring gyro (0.0035 deg/sqrt-hour) and that of a MEMS accelerometer (2.25 deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if done in a chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no export restrictions by using such an array? Or will ITAR or the like be slapped on such a device once its accuracy is published in a brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial Sensors with a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to sync
up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If the platform has rotational acceleration there would be a difference in the acceleration on each device depending on it's distance from the center. That would keep them out of sync.

Would it? If they're on the same platform, the coupling is the same
no matter what additional force is on them (superposition).


Krw happens to be wrong here. The force imposed by rotational acceleration does vary in proportion to the distance from the centre of rotation, as Rick pointed out, and krw seems to have failed to comprehend.

--
Bill Sloman, Sydney


Guest

Mon Feb 11, 2019 4:45 am   



js5071921_at_gmail.com wrote in
news:d2ea4c11-7ad7-411f-8d1e-ea2edbb53625_at_googlegroups.com:

Quote:
On Sunday, February 10, 2019 at 6:28:43 PM UTC+2,
gnuarm.del...@gmail.com wrote:
On Sunday, February 10, 2019 at 10:57:41 AM UTC-5, John Larkin
wrote:
On Sat, 9 Feb 2019 21:58:58 -0800 (PST), JS
wrote:

On Sunday, February 10, 2019 at 1:13:07 AM UTC+2, John Larkin
wrote:
On Sat, 9 Feb 2019 12:47:28 -0800 (PST),
wrote:

Hi all,

Given that the random error in a sample is proportional to
1/sqrt(s
ample size), does having many accelerometers and then averaging
their output therefore reduce their overall error?

So would it be worthwhile to have say 100 or 1000 cheap
acceleromet
ers rather than one expensive one like a laser ring gyro?

Thanks.

sqrt(1000) is only 32. I'd expect the ring gyro to be vastly
better than a cheap MEMS or some such.


--

John Larkin Highland Technology, Inc

lunatic fringe electronics

OK I did the sums. Based on the random walk of a laser ring
gyro (0.00
35 deg/sqrt-hour) and that of a MEMS accelerometer (2.25
deg/sqrt-hour) [1], you need about 400k MEMS accelerometers to
approach the accuracy of a laser ring gyro.

It sounds like a lot of components to solder together but if
done in a
chip fab, it should be possible.

Is it possible to make a commercial accelerometer with no
export restr
ictions by using such an array? Or will ITAR or the like be
slapped on such a device once its accuracy is published in a
brochure?

Refs:
[1] Honeywell GG1320AN Digital Laser Gyro brochure
[2] Error and Performance Analysis of MEMS-based Inertial
Sensors with
a Low-Cost GPS Receiver. Park, M & Gao, Y. [2008] Sensors Vol 8

If the MEMS parts use vibrating cantilevers, they would want to
sync up. I don't know if that is good or bad.

They might "want" to sync up, but I'm not sure they would. If
the platfo
rm has rotational acceleration there would be a difference in the
acceleration on each device depending on it's distance from the
center. That would keep them out of sync.

Rick C.

If the body is rigid then s = r*theta, and you can take the time
derivative twice for velocity and then acceleration.

Problem is that real life is not so easy. A long slender missile
body for example would have very different transient accelerations
at the tip and the tail. After you double integrate the
accelerations to get displacements, over a distance of say 10 -
1000km range, your CEP becomes wide as the side of a barn.

The best thing really for such applications is to either locate
all your cheap accelerometers into a small, rigid box, or use a
more niche but higher accuracy accelerometer like a RLG.


My units used a GE IRU (Inertial Reference Unit) They are about a 7
inch cube and run about 50k each in quantity. Singles usually cost
you about $100k each. We used them to keep a satellite dish aimed
while the base it was mounted on was in motion (mobile sat comms).

<http://pdf.aeroexpo.online/pdf/ge-aviation/9181-series-inertial-
reference-unit/169836-10741.html>
3 to 5 mil accuracy.

Goto page Previous  1, 2, 3, 4  Next

elektroda.net NewsGroups Forum Index - Electronics Design - Using many cheap accelerometers to reduce error

Ask a question - edaboard.com

Arabic version Bulgarian version Catalan version Czech version Danish version German version Greek version English version Spanish version Finnish version French version Hindi version Croatian version Indonesian version Italian version Hebrew version Japanese version Korean version Lithuanian version Latvian version Dutch version Norwegian version Polish version Portuguese version Romanian version Russian version Slovak version Slovenian version Serbian version Swedish version Tagalog version Ukrainian version Vietnamese version Chinese version Turkish version
EDAboard.com map