Opencpn backup of multiple fixes using Kalman filter

[PjotrC]

Originally Posted by bcn

So how to get this in OpenCPN?

Hubert,

The simulator outputs to COM2 which is a virtual port, in my case set up with VSPE Virtual Serial Ports Emulator. With practice, it takes less than a minute to have it all running. Maybe some scripting can even speed this up.

FlyWithCE GPS Simulator is the simplest one I did find around. It is not free in general and has some annoying behaviour, but is quite usable and useful.

The drawback of having this solution not integrated in OCPN is that the OCPN logkeeping is not aware of the fix quality. edit: this could be easily included by setting number of GPS satellites to - say - 1.

Some chartplotters, like MaxSea (if I remember correctly) switch automatically to DR mode, assuming last COG and SOG received. I am not sure if this automation is really needed. Perhaps it might even be harmful... Upon losing GPS fix the navigator should obviously put on a different thinking cap...

What Rick suggests sounds more like another Chartwork application, taking a number of fixes and producing another, "better" fix ... could be a plugin of course.

Piotr

[bcn]

Originally Posted by PjotrC

Hubert,

Some chartplotters, like MaxSea (if I remember correctly) switch automatically to DR mode, assuming last COG and SOG received. I am not sure if this automation is really needed. Perhaps it might even be harmful... Upon losing GPS fix the navigator should obviously put on a different thinking cap...
Piotr

Yes, that would be my approach as well
- on GPS loss alert
- give the option to go DR with inputs outlined by you
- until GPS returns (or the ship did run aground in the meantime)

[rgleason]

I like what you guys are suggesting with a staged approach:
1. Simple DR connection mode to outside software.
2. FS#1584 - Implement Dead Reckoning as fall back at GPS outage
3. Plugin for Kalman Filtered DR (similar to celestial navigation perhaps, but more integrated.
4. Plugin for Nmea Inertial devices.

I suppose this is pie in the sky stuff!
but Thanks!

[petter5]

Originally Posted by bcn

Yes, that would be my approach as well
- on GPS loss alert
- give the option to go DR with inputs outlined by you
- until GPS returns (or the ship did run aground in the meantime)

some position sensors like Digial Yatch Dualnav Home will give you both GPS and GLONASS nmea data from the same sensor (and also Galileio when available in future) This is in reality multiple fixes in the same nmea stream that cover different navigational satellite systems from the same sensor. It would be a very dependable backup solution to make OpenCpn read also glonass NMEA0183 sentenses and multiplex these sentenses using a Kalman filter, or define a primary and secondary sensor and have a fallback or diff function. The dualnav sensor does even cost less than garmin/simrad and similar marine gps sensors.

[dacust]

My thoughts are to have a utility/PI/something that would take any information I gave to it and tried to narrow down where I am. GPS, inertial, sextant, DR guess, manual speed log, typed in magnetic compass heading, degrees bearing to a fixed point I have on a chart, time of day that the sun reaches zenith.

Kinda ties in to another subject. One old GPS I had showed a circle around my position as big as the potential error. Using the above, if it was sextant, then maybe you put in your confidence level and it then draws a 2 mile ring, or whatever.

To really go further, it could keep track of each sextant sight you take and correlate it with pitch/yaw/frequency of boat movement/etc. and come up with it's own estimate if how accurate your sights usually are depending on sea state.

-dan

[seandepagnier]

I think a wind speed sensor would make more sense than inertial sensors for gps outage, however a magnetometer would allow detecting wind shifts. A water speed sensor would help as well.

That is because using accelerometers to calculate position has just too many errors and isn't accurate at all. You need the accelerometers and gyroscopes to tilt-compensate the magnetometers.

Maybe you can use the accelerometer data to determine changing seastates, and adjust course/speed.. but this is complex. Maybe it's possible with enough backlog when the gps was working.

If the boat is sailing along and the wind and conditions remain the same, then the speed is likely the same.


I don't think the kalman filter should be used here. You are thinking of the case of using gps _and_ accelerometers at the same time, and the kalman filter can merge them into much more accurate position data than with gps alone. Without gps, I don't know what kalman filter you mean, unless it's for merging inertial sensors?

[rgleason]

Sean it appears our discussion was pretty futile. Ok what filter should we be talking about that can take multiple info and estimate position?

Sent from my SCH-I545 using Cruisers Sailing Forum mobile app

[seandepagnier]

kalman filter can do this, but it needs some sort of position feedback or it will drift from integration errors.

So in theory, you could use all sorts of things, celestial fix, magnetic latitude/field strength, land sightings, directional antenna for am radio, fog horns, visually seeing reflected swells from some islands far away, etc etc...

You cannot have only inertial sensors and wind/water speed sensors and no data related to position on the earth or over time you just won't know where you are because integration errors accumulate over time.

[rgleason]

By the way samsung s4 and other phones have inertial sensors acceleration, mag field mag heading, altitude pitch roll etc. To bad there isnt a way to use this in opencpn somehow even if sean says its not useful for position. Maybe useful for apilot steering?

Sent from my SCH-I545 using Cruisers Sailing Forum mobile app

[rgleason]

These devices are also inside tablets...

Sent from my SCH-I545 using Cruisers Sailing Forum mobile app

[rgleason]

Aren't we able to configure a second GPS Sensor?
Are we then able to call it a backup GPS sensor? to have it set on "standby" backup mode?

[bcn]

Different COM ports with different priorities - priorities have been invented for that purpose -will act as a back-up.

If you have the data coming from an UDP stream with two GPS you might see some jumping (usually on the last digits) and when one dies you have still the second one.
Here you might play with the sender ID, the first two characters of the NMEA0183 message ($GPRMC where GP is the sender-ID and RMC identifies the payload).
But usually you will not be able to define the sender ID of a given device.

In NMEA2000 you can identify instances of sensors and same for Signal K.

[rgleason]

Small Positioning Chip
TIMU, or timing and inertial measurement unit, was created by DARPA-funded researchers at the University of Michigan, as part of their Micro-PNT project.

No Gps The TIMU, DARPA recently announced, performs the three functions necessary for navigation, simultaneously measuring orientation, acceleration and time.

The chip has a six-axis inertial measurement unit made of three gyroscopes and three accelerometers, along with a master clock. It’s built from groundbreaking designs from DARPA’s Micro-Technology for Positioning, Navigation and Timing (Micro-PNT) program, launched in January 2010 with the goal of developing microscale clocks and inertial sensors. Six microfabricated layers make up the TIMU, each one 50 microns thick (the thickness of a human hair) and each performing a different function, DARPA said. Altogether, the whole package takes up 10 cubic millimeters. (Uses 3d printing technology)
".. small enough and should be robust enough for applications (when GPS is unavailable or limited for a short period of time) such as personnel tracking, handheld navigation, small diameter munitions and small airborne platforms.”

DARPA Beyond GPS: 5 Next-Generation Technologies for Positioning, Navigation & Timing (PNT)

DARPA Reducing Tics in the Tocks of Atomic Clocks

DARPA’s Micro-Technology for Positioning, Navigation and Timing (Micro-PNT) program

DARPA Subs Positioning System for Deep Ocean Navigation(POSYDON)

Furuno has some interesting commercial offerings
GPS/GLONASS/Galileo/SBAS/QZSS Receivers for Positioning and Timing markets
"Positioning and Timing solutions, enabled to receive signals of almost all GNSS Systems available. We also support Dead Reckoning (DR), High Positioning Rate (e.g. 10Hz) for positioning. For timing solutions, our products are also capable of 1PPS output as well as Clock output functionality(configurable to, e.g. 10MHz) and data output in NMEA or M12 (Motorola) Binary protocol."

Furuno GV-87 Dead Reckoning + Concurrent Multi-GNSS Receiver Module GV-87 is the concurrent Multi-GNSS Receiver Module supporting Dead Reckoning (DR) function, enabling high accuracy positioning with Multi-GNSS and also positioning in environments where no satellite signals are received. -Sample model available.
Search on a video: "Performance of FURUNO Dead Reckoning Multi-GNSS Receiver Module GV-87"

Furuno GV-86 New Small Package, High Sensitivity DR (Dead Reckoning) + GPS Module GV-86 is the high sensitivity GPS Receiver Module supporting Dead Reckoning (DR) function enabling positioning in environments where no satellite signals are received -Sample available.


What is a GPS/Multi-GNSS Disciplined Oscillator


Multi-GNSS (Multi-frequency GNSS)


These devices are probably quite expensive, but perhaps someone knows more about this technology. There is no doubt that integrated high performance DR chips are in our future, in a very basic inaccurate form they already exist in your cell phone.


There may be some early tinkerers and technos who have already created and are using similar solutions.

[Gnomon]

Multi-GNSS

Rick,
just installed a USB-2 GNSS Receiver with u-blox 8 chipset, less than EUR 100.
Inside window of lifingroom, beneath balcony and overhead Powerline, more than 20 satelittes in view, about 10 used. Windows 10. Simple Installation.

Navilock Home (also in English)
www.u-blox.com

In the Moment GPS and GLONASS, Galileo, BeiDue and QZSS.
SBAS: EGNOS, WASS, MSAS, GAGAN, and SDCM will follow.
Horizontal Error 5 to 10 m, after 7500 samples.

Wolfgang

[rgleason]

Wolfgang. Thank you for responding. Sure wish the gps and gyrocompass could be joined with DR and greater accuracy with nmea0183 or nmea2000. Maybe it is available now!