I guess that makes me a "nobody" then, since I have had the unfortunate pleasure of inheriting a project at work that attempts to shoehorn a vxworks project into linux via v2lin.
Given enough time, I would love to chuck v2lin out altogether and access pthreads directly.
Does anybody have any tips for surgically removing v2lin from legacy code?
Firstly, with regular GPS you already have more than one clock - one on your receiver, and one on each of the satellites. You can directly solve for the receiver clock bias by taking measurements to an extra satellite, hence the need to track 4 satellites for a three-dimensional position fix because of the four unknowns ( X, Y, Z, and time.)
Secondly, while GPS and GLONASS use different terrestrial reference frames, there exists a well defined transform between the WGS-84 used by GPS and the PZ-90 used in GLONASS.
Finally, in a combined GPS/GLONASS receiver it is not optimal to calculate a separate position solution for each constellation. If you track a few more satellites, you can solve directly for the clock offset between the two navigation systems and treat the range measurements as if they were all from one giant 60 satellite constellation. This actually gives you much better satellite geometry, and is often more accurate than any single navigation system on its own.
There is much research being done on the effects of combined constellations with GPS,GLONASS, Galileo and the Chinese Compass system.
Also, the article seems to imply that GPS is used for drawing pictures:
"work together to provide more accurate images and information"
"would be able to create a more accurate picture especially in areas where reception is weak"
Of your 5 satellites visible, i would be betting one of those was a WAAS, so you can only really count the 4 sats.
The critical thing is the Dilution of Precision (DOP). Your accuracy on the ground is directly proportional to the DOP.
Basically, (thinking two-dimensionally here for the purpose of the exercise) you can think of this value as the area of a polygon with your receiver on one point, and the gps satellites making up the other points. Having more satellites (ie. combined constellation) along the short edge of a long skinny polygon will not significantly increase the area.
Same goes for your DOP. I have done some experiments with combined GPS glonass solutions, and if you are in a crummy environment like down a trench or in a built up city, the combined solution accuracy is not always better than GPS alone.
That being said, having more sats in the sky DOES increase your availability of a position because you reduce your chances of losing lock on all your sats at the same time when moving through a built up environment.
In practice the next-generation GNSS receivers will be dual-constellation, and will use BOTH of the GPS and Gailieo ranging signals.
The combined constellation of 54 satellites (24 GPS + 30 Galileo) will have advantages in terms of satellite availability ( number of satellites above a certain elevation angle), and the geometry (more satellites to chose from will allow the choise of a better dilution of precision, therefore better position accuracy for given range measurement errors).
Ultimately this will benefit consumers, especially in built up areas, more than either system on its own.
My lab was involved a few years back with an high performance computing experiment with self-healing FPGA's after exposure to ionising radiation.
They flew some off the shelf (non radiation hardened) FPGAs on the FedSAT-1 spacecraft.
I was involved with a different (GPS) payload, but i believe the HPCE payload was able to successfully self-diagnose and correct single gate errors on the chip. (http://www.crcss.qut.edu.au/comp/hpce.pdf)
Slashdot Mirror
I guess that makes me a "nobody" then, since I have had the unfortunate pleasure of inheriting a project at work that attempts to shoehorn a vxworks project into linux via v2lin.
Given enough time, I would love to chuck v2lin out altogether and access pthreads directly.
Does anybody have any tips for surgically removing v2lin from legacy code?
Firstly, with regular GPS you already have more than one clock - one on your receiver, and one on each of the satellites. You can directly solve for the receiver clock bias by taking measurements to an extra satellite, hence the need to track 4 satellites for a three-dimensional position fix because of the four unknowns ( X, Y, Z, and time.)
Secondly, while GPS and GLONASS use different terrestrial reference frames, there exists a well defined transform between the WGS-84 used by GPS and the PZ-90 used in GLONASS.
Finally, in a combined GPS/GLONASS receiver it is not optimal to calculate a separate position solution for each constellation. If you track a few more satellites, you can solve directly for the clock offset between the two navigation systems and treat the range measurements as if they were all from one giant 60 satellite constellation. This actually gives you much better satellite geometry, and is often more accurate than any single navigation system on its own.
There is much research being done on the effects of combined constellations with GPS,GLONASS, Galileo and the Chinese Compass system.
Also, the article seems to imply that GPS is used for drawing pictures:
:-) /
"work together to provide more accurate images and information"
"would be able to create a more accurate picture especially in areas where reception is weak"
So, farmers ploughing profanities in their fields will be able to use better fonts now.
http://www.theregister.co.uk/2006/05/31/huge_word
Of your 5 satellites visible, i would be betting one of those was a WAAS, so you can only really count the 4 sats. The critical thing is the Dilution of Precision (DOP). Your accuracy on the ground is directly proportional to the DOP. Basically, (thinking two-dimensionally here for the purpose of the exercise) you can think of this value as the area of a polygon with your receiver on one point, and the gps satellites making up the other points. Having more satellites (ie. combined constellation) along the short edge of a long skinny polygon will not significantly increase the area. Same goes for your DOP. I have done some experiments with combined GPS glonass solutions, and if you are in a crummy environment like down a trench or in a built up city, the combined solution accuracy is not always better than GPS alone. That being said, having more sats in the sky DOES increase your availability of a position because you reduce your chances of losing lock on all your sats at the same time when moving through a built up environment.
Professor Werner Enderle from QUT has done extensive modelling and simulation on galileo performance for LEO missions. Have a look at http://www.gmat.unsw.edu.au/wang/jgps/v2n2/v2n2For umB.pdf
In practice the next-generation GNSS receivers will be dual-constellation, and will use BOTH of the GPS and Gailieo ranging signals. The combined constellation of 54 satellites (24 GPS + 30 Galileo) will have advantages in terms of satellite availability ( number of satellites above a certain elevation angle), and the geometry (more satellites to chose from will allow the choise of a better dilution of precision, therefore better position accuracy for given range measurement errors). Ultimately this will benefit consumers, especially in built up areas, more than either system on its own.
My lab was involved a few years back with an high performance computing experiment with self-healing FPGA's after exposure to ionising radiation.
They flew some off the shelf (non radiation hardened) FPGAs on the FedSAT-1 spacecraft.
I was involved with a different (GPS) payload, but i believe the HPCE payload was able to successfully self-diagnose and correct single gate errors on the chip. (http://www.crcss.qut.edu.au/comp/hpce.pdf)