[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PSUBS-MAILIST] Navigation
Hi Sean
thanks for providing a back ground on INS and you experiences attempting
to implement such as system.
I would make distinction between inertial navigation and dead reckoning
navigation. Dead reckoning would involve starting at known position,
then traveling on fixed heading at a certain speed for a given time,
then using: Distance = Velocity * Time to calculate(/estimate) the
distance traveled along the heading to estimate the new position.
Inertial navigation would involve starting at know position and recording
inputs to sensor, such as accelerometers and gyros, aggregating (and
filter) the data to estimate the new position. Both systems are
susceptible to similar but slightly different errors. They can also
be combined to provide a more accurate estimation of the new position.
Another thing to note when dealing with INS/IMUs is that flux gate
compasses are not really inertial measurement devices, but measure
terrestrial magnetic field and use the readings to augment the INS
results.
I've not really had a chance to do any work on my INS for a few months,
I really wish I'd gotten to show it off at the convention. Although
it doesn't work properly, in the sense that it can not accurately
provide position, it does provide an accurate estimation of the state
of the vessel. Pitch, roll, heading and the ability to tell in which
directions forces are operating on the vessel (can you tell if we are
accelerating slowly, or if we are snagged on something?). I was about
20seconds into a dive on board the Bionic Guppy when I realized that
this stuff was already useful for psubs.
The sensor unit I start out with is the Ocean Server OS5000:
http://www.oceanserver-store.com/osevkitcowid.html
After playing around with it for a while I started looking for something
more accurate/sensitive/reliable. In the end I found a fiber optic gyro
for KVH, which is accurate to around 0.01 degrees, I can measure the earth's
rotation with this unit. I am still looking to collect more sensors with
a similar accuracy.
My plans for the inertial nav system I have been working on is to:
- convert the math/rotation engine from Euclidean geometry to use quadronations.
- expand the Kalman filter code to handle more sensors with it being at lot
of work to add sensors.
- drop the USB ADC I am using to read data from the fiber optic gyro and
using something with better timing.
- add more "cheap" IMU sensors
- add physics model to the filter (if the sub turns 90degs in 1 sec, then
there is a bad sensor reading).
- add non-IMU environment sensors, water speed, depth, flux gate compass.
- add non-IMU system sensors and inputs (rudder angle, main motor power, etc.)
- add more "expensive" IMU sensors
- use DGSP rather than GSP to get initial position.
- combine with active sonar doing feature recognition.
That is about 5 masters degrees and 2 PhDs there. :-)
If I ever get that far then I want to play around with some machine learn
to apply the sensor data from optimum accuracy. For example, if we have
5 heading gyros and a flux gate compass, then the flux gate compass starts
turning (due to iron in the area), but the gyros do not, then the system can decide
to ignore the compass readings (and just deal with gyro readings and gyro
drift for a while), once the gyros and compass readings converge again
(the sensor is not longer effected by the iron), then deviation between the
compass and gyro can be used to estimate the gyro drift that occurred while
the compass reading was discarded, this can be applied the previous data
to improve the accuracy of the position during the compass outage.
But, back to the present and reality, I think justing something like the
OS5000 device + plus something to display the results on very useful to
most psubs. I don't know how well the flux gate compass will work, but
this particular device is designed to be used by ROVs (mounted far away
from thrusters), so I think it should work on psub. I does need mounted
outside of the hull if you want to use the compass feature.
Cheers!
Ian.
-----Original Message-----
>From: "Sean T. Stevenson" <cast55@telus.net>
>Sent: Feb 12, 2011 7:54 PM
>To: personal_submersibles@psubs.org
>Subject: Re: [PSUBS-MAILIST] Navigation
>
>Yes - as long as you know the system state when the IMU is activated,
>then you can use the IMU output to determine the current state. The
>IMU only measures acceleration directly, but if your ship is floating
>with the current, you can get both the initial position and the initial
>velocity from the GPS. Using both of those values as the startup state,
>the IMU can provide corrections from then on. Whenever you have a valid
>GPS fix, you would re-initialize the internal velocity and position to
>that reported by the GPS in order to eliminate the accumulated errors,
>"resetting the clock", so to speak. (Incidentally, GPS is actually more
>accurate in velocity than it is in position. This makes sense if you
>think about how a GPS receiver actually works.) Similarly, using a
>Doppler sonar when near the bottom will solve half the problem - you
>can't get absolute position information, but you can zero the velocity
>error.
>
>-Sean
>
>
>On 12/02/2011 5:50 PM, Alan James wrote:
>> Thanks Sean,
>> So if the mother ship isn't anchored & is flowing in a current the
>> unit won't pick this up, as the submarine upon
>> launching won't be accelerating.
>> If the mother ship was anchored in a current the unit would need
>> turning on prior to the sub being
>> released into the current to pick up that movement.
>> There was an additional unit with gyro for $100- US
>> http://www.cytron.com.my/viewProduct.php?pid=EAUJAwYiADILFCsTIhw4HkqL6/mfx3ZbyBEjcLFbeI4=
>> Regards Alan
>>
>
************************************************************************
************************************************************************
************************************************************************
The personal submersibles mailing list complies with the US Federal
CAN-SPAM Act of 2003. Your email address appears in our database
because either you, or someone you know, requested you receive messages
from our organization.
If you want to be removed from this mailing list simply click on the
link below or send a blank email message to:
removeme-personal_submersibles@psubs.org
Removal of your email address from this mailing list occurs by an
automated process and should be complete within five minutes of
our server receiving your request.
PSUBS.ORG
PO Box 53
Weare, NH 03281
603-529-1100
************************************************************************
************************************************************************
************************************************************************