Vladimir Vassilevsky <nospam@nowhere.com> writes:> On 7/22/2013 11:21 AM, Randy Yates wrote: >> Randy Yates <yates@digitalsignallabs.com> writes: >>> [...] >>> What I don't understand is why, no matter how inaccurate a sensor >>> might be, you couldn't calibrate it out to some degree. >> >> Clarification/restatement of question: >> >> What I don't understand is why, no matter how inaccurate a sensor might >> be, you couldn't calibrate out the offset to some degree. > > You are trying to calibrate wooden rule to sub-micron accuracy. > Good luck.So how does reality differ from the model I presented? Is offset a function of orientation? Of temperature? Of moon phase? Is the offset non-systematic (even at one orientation and one temperature and a full moon)? What???!? Just curious at this point. -- Randy Yates Digital Signal Labs http://www.digitalsignallabs.com
IMU Sensor Calibration Tricks?
Started by ●July 18, 2013
Reply by ●July 22, 20132013-07-22
Reply by ●July 22, 20132013-07-22
Vladimir Vassilevsky <nospam@nowhere.com> wrote:> On 7/22/2013 11:21 AM, Randy Yates wrote:(snip)>> What I don't understand is why, no matter how inaccurate >> a sensor might be, you couldn't calibrate out the offset >> to some degree.To some degree, yes. But to what degree?> You are trying to calibrate wooden rule to sub-micron accuracy. > Good luck.Note that a wooden ruler will change with temperature and humidity. If you keep those constant, and wait until it reaches equilibrium, it might stay pretty constant. The data sheet for the sensor mentions package stress in soldering. If the thermal expansion of the PC board is different, there will also be a thermal component to the package stress. Also, it might take some time to reach equilibrium. And maybe more reasons that that. I didn't notice in the data sheet if it said how good the calibration was supposed to be. -- glen
Reply by ●July 22, 20132013-07-22
On Mon, 22 Jul 2013 09:21:30 -0700 (PDT), clay@claysturner.com wrote:>On Monday, July 22, 2013 12:11:50 PM UTC-4, Vladimir Vassilevsky wrote: >> On 7/22/2013 11:00 AM, jacobfenton wrote: >> >> >> >> > They rely heavily on GPS, if they flew without GPS, I am pretty sure they >> >> > would not navigate so well over a long duration. >> >> >> >> Dead reconing using magnetic bearing is viable, too. >> >> >> >> VLV > >Red foxes use magnetic fields to improve their targeting of rodents under the snow. > >See http://rsbl.royalsocietypublishing.org/content/early/2011/01/06/rsbl.2010.1145.full > >I don't think they use MEMS.But they do have good inertial sensing. So do cats. ;)> >Clay > > > > >Eric Jacobsen Anchor Hill Communications http://www.anchorhill.com
Reply by ●July 22, 20132013-07-22
jacobfenton <51114@dsprelated> wrote: (snip on MEMS sensors)> They rely heavily on GPS, if they flew without GPS, I am pretty > sure they would not navigate so well over a long duration.They could use GPS for long range, and MEMS for short range (short time scale) sensing. Maybe it doesn't matter for rockets.> There are tatical grade MEMS, used for small rockets, but those > flights are short, so the drift and issues with MEMS probably > are not as apparent. Navigation grade MEMS, would mean no > GPS updates, flying free inertial only, its just not there yet.Portable car navigation systems might do it using only GPS, but built-in ones can use the odometer signal for distance, and a gyro to sense rotation, with the GPS signal to correct for drift in the long term. One we have has a coriolis force based gyro sensor, as far as I know, not MEMS. (I don't know what the MEMS sensors are like.) Consider a bending mode piezoelectric material with a sine voltage, moving a mass back and forth. Now, using another piezoelectric device measure the signal on a perpendicular axis. The magnitude and phase give the rotation rate. -- glen
Reply by ●July 22, 20132013-07-22
On Mon, 22 Jul 2013 11:00:15 -0500, "jacobfenton" <51114@dsprelated> wrote:>>On Mon, 22 Jul 2013 10:01:48 -0500, "jacobfenton" <51114@dsprelated> >>wrote: >> >>>>Vladimir Vassilevsky <nospam@nowhere.com> writes: >>>> >>>>> On 7/20/2013 6:09 PM, Randy Yates wrote: >>>>>> eric.jacobsen@ieee.org (Eric Jacobsen) writes: >>>>>>> [...] >>>>>>> Naturally, increasing precision requires exponential effort/cost, >but >>>>>>> Randy hasn't indicated (that I've noticed) how much precision he >>>>>>> needs. >>>>>> >>>>>> That's cause Randy has no freakin' idea! >>>>>> >>>>>> Well, I can make a stab at offset: less than 0.0005 feet/sec^2. >>>>> >>>>> Forget about that. With sensors of MEMS class, your very very very >>>>> best bet would be about 1 cm/s^2. >>>> >>>>So you're saying that even if the sensor output signal were to be >>>>"calibrated" to "exactly" 0 offset, the sensor itself would have an >>>>"inherent" offset of 1 cm/s^2 because it is MEMS? >>>>-- >>>>Randy Yates >>>>Digital Signal Labs >>>>http://www.digitalsignallabs.com >>>> >>> >>>Vlad is right, there is a reason there are no Navigation grade MEMS yet. > >> >>What do they use in the tiny UAVs that are autonomous or >>semi-autonomous? Many of those are too small to have anything bigger >>than MEMs technology. Likewise many of the guided amateur rocket >>launches. >> >> >> >> >> >>Eric Jacobsen >>Anchor Hill Communications >>http://www.anchorhill.com >> > >They rely heavily on GPS, if they flew without GPS, I am pretty sure they >would not navigate so well over a long duration. There are tatical grade >MEMS, used for small rockets, but those flights are short, so the drift and >issues with MEMS probably are not as apparent. Navigation grade MEMS, would >mean no GPS updates, flying free inertial only, its just not there yet. >Maybe some day it will.Even the expensive strap-down IMUs can't be relied on forever and need updating (often by GPS), so I don't see a difference other than in degree, which is expected. My point was that there are navigation tasks being done with MEMs used for IMUs. But I think Randy's application isn't navigation, anyway, so it's a bit moot.> >_____________________________ >Posted through www.DSPRelated.comEric Jacobsen Anchor Hill Communications http://www.anchorhill.com
Reply by ●July 23, 20132013-07-23
On Monday, July 22, 2013 2:07:05 PM UTC-4, Eric Jacobsen wrote:> On Mon, 22 Jul 2013 11:00:15 -0500, "jacobfenton" <51114@dsprelated> > > wrote: > > > > >>On Mon, 22 Jul 2013 10:01:48 -0500, "jacobfenton" <51114@dsprelated> > > >>wrote: > > >> > > >>>>Vladimir Vassilevsky <nospam@nowhere.com> writes: > > >>>> > > >>>>> On 7/20/2013 6:09 PM, Randy Yates wrote: > > >>>>>> eric.jacobsen@ieee.org (Eric Jacobsen) writes: > > >>>>>>> [...] > > >>>>>>> Naturally, increasing precision requires exponential effort/cost, > > >but > > >>>>>>> Randy hasn't indicated (that I've noticed) how much precision he > > >>>>>>> needs. > > >>>>>> > > >>>>>> That's cause Randy has no freakin' idea! > > >>>>>> > > >>>>>> Well, I can make a stab at offset: less than 0.0005 feet/sec^2. > > >>>>> > > >>>>> Forget about that. With sensors of MEMS class, your very very very > > >>>>> best bet would be about 1 cm/s^2. > > >>>> > > >>>>So you're saying that even if the sensor output signal were to be > > >>>>"calibrated" to "exactly" 0 offset, the sensor itself would have an > > >>>>"inherent" offset of 1 cm/s^2 because it is MEMS? > > >>>>-- > > >>>>Randy Yates > > >>>>Digital Signal Labs > > >>>>http://www.digitalsignallabs.com > > >>>> > > >>> > > >>>Vlad is right, there is a reason there are no Navigation grade MEMS yet. > > > > > >> > > >>What do they use in the tiny UAVs that are autonomous or > > >>semi-autonomous? Many of those are too small to have anything bigger > > >>than MEMs technology. Likewise many of the guided amateur rocket > > >>launches. > > >> > > >> > > >> > > >> > > >> > > >>Eric Jacobsen > > >>Anchor Hill Communications > > >>http://www.anchorhill.com > > >> > > > > > >They rely heavily on GPS, if they flew without GPS, I am pretty sure they > > >would not navigate so well over a long duration. There are tatical grade > > >MEMS, used for small rockets, but those flights are short, so the drift and > > >issues with MEMS probably are not as apparent. Navigation grade MEMS, would > > >mean no GPS updates, flying free inertial only, its just not there yet. > > >Maybe some day it will. > > > > Even the expensive strap-down IMUs can't be relied on forever and need > > updating (often by GPS), so I don't see a difference other than in > > degree, which is expected. My point was that there are navigation > > tasks being done with MEMs used for IMUs. >A number of UAVs have synthetic aperture radar capability - to achieve the positional accuracy required they blend the IMU and GPS data via a Kalman Filter. On larger aircraft the blend the strapdown IMU, the aircraft INS (inertial navigation system) and GPS. Cheers, Dave
