How accurate would you like your GPS?

[Topcat47]

At sea, +_ 500 yds is OK for most purposes. My Garmin has placed my boat in the middle of Gosport High Street on a number of occasions, which is quite worrying, except I don't really need to navigate by electronic means in Haslar Creek. If I can't see Haslar Marina from my berth, I don't sail. GPS or No GPS, you won't find me entering the inner swashway in poor viz as there aren't enough contour lines to run by depth readings. It's important to make seamanlike decisions, GPS or no GPS. I rather liked John Goode's assertion that if I navigate across channel from Pompey to Cherbourg, without a "fix" I have a circle of uncertainty of around my estimated position of about 2 miles. I think it came from "crossing with confidence" but I can't find my old copy or remember the method of calculation. Luckily, thanks to GPS, I'm usually better located than that these days and as long as I can get a fix on Barfleur with my hand bearing compass I"m reasonably confident even at night.

 

[grumpy_o_g]

Aircraft have been able to fly GPS approaches for some time albeit with higher minima, but an improvement in the accuracy will be appreciated. Inevitably there has always been a danger relying in GPS and especially the danger of local interference or any other issues.

The main issue with using GPS for IFR approaches was the altitude information - always a weak spot with GPS compared to lateral information. It's now plenty accurate enough to fly down to minimums at which point the Mk 1 eyeball takes over anyway. Still quite a way from being accurate enough to use for auto-land though. Oddly enough even ILS-based auto-land needs a scatter factor built in for lateral position as it's so accurate every a/c would hit exactly the same spot and soon damage the runway. I'd be interested to know how much this improves the vertical accuracy. I'm guessing it's not certified for aviation yet as that usually takes quite a while - there's a quite a few EGNOS systems around already (at reasonable prices by aviation standards too) and these are far more accurate vertically but I'm not aware of any plans to extend VNAV (vertical nav by GNSS) down to zero.

Apologies if I'm teaching grandmother suck eggs (where does that phrase come from?).

 

[RAI]

The main issue with using GPS for IFR approaches was the altitude information - always a weak spot with GPS compared to lateral information. It's now plenty accurate enough to fly down to minimums at which point the Mk 1 eyeball takes over anyway. Still quite a way from being accurate enough to use for auto-land though. Oddly enough even ILS-based auto-land needs a scatter factor built in for lateral position as it's so accurate every a/c would hit exactly the same spot and soon damage the runway. I'd be interested to know how much this improves the vertical accuracy. I'm guessing it's not certified for aviation yet as that usually takes quite a while - there's a quite a few EGNOS systems around already (at reasonable prices by aviation standards too) and these are far more accurate vertically but I'm not aware of any plans to extend VNAV (vertical nav by GNSS) down to zero.

Apologies if I'm teaching grandmother suck eggs (where does that phrase come from?).

I think approaches in Cat I conditions are available to air transport both with differential GPS (GBAS) and with slightly different rules with EGNOS (SBAS). Cat II/III weather operation is another thing. ICAO were still working on the GLS standards ten years ago when I left the game. Integrity is the problem. It will be a lot easier with dual frequency receivers, now that both the signals and receivers are available. Dear old ILS keeps on being supported, warts and all. After 30 years of standardization the MLS system got abandoned, maybe GBAS/SBAS will go the same way.

 

[prv]

I think approaches in Cat I conditions are available to air transport both with differential GPS (GBAS) and with slightly different rules with EGNOS (SBAS). Cat II/III weather operation is another thing. ICAO were still working on the GLS standards ten years ago when I left the game. Integrity is the problem. It will be a lot easier with dual frequency receivers, now that both the signals and receivers are available. Dear old ILS keeps on being supported, warts and all. After 30 years of standardization the MLS system got abandoned, maybe GBAS/SBAS will go the same way.

This reminds me of the accident reports on two of the Watchkeeper drone crashes

They have both a kind of portable ILS system, and the ability to land on GPS alone if the portable beacon isn't set up. Both systems operate together with laser altimeters that don't work well in cloud. What they don't have is a simple weight-on-wheels switch to recognise when they've landed; they try to simulate it with accelerometers. A bit of turbulence over the threshold, in cloud, with the portable beacon unserviceable, and it decided it had landed when still 60 feet above the runway.

The first thing it does after landing is pitch down hard to put weight on the nosewheel for steering...

Pete

 

[savageseadog]

Knowing sea level, ie the tidal height might be useful. Currently GPS with WAAS heights aren't accurate enough to be useful.

 

[Tomaret]

I've not sailed in Sweden, but am looking forward to doing so in the future. At distances less than 50m, probably 500m, I use the Mark I eyeball which I find far more accurate for the task in hand.

When you are looking for the west entrance to Cherbourg Harbour in less than 200m visibility as we were last Saturday, with no radar on board, Mark 1 eyeball alone isn’t enough. Knowing that the chartplotter is showing where you are accurate to a couple of boat lengths is very helpful. Only when we were actually in the marina could we see where we were. It was also helpful that as we knew where we were we could concentrate our efforts on looking out for where other boats might be. Fortunately nobody came within sight of us. The experience has persuaded the skipper that even if he can’t afford radar an AIS receiver is worth investing in. It certainly engenders respect for those that have gone before us without radar, ais or chartplotter.

 

[Joker]

The Mark 1 eyeball often doesn't spot ledges, or places where the depth goes from 20m to 1m in about a boat's length.

 

[TQA]

The Mark 1 eyeball often doesn't spot ledges, or places where the depth goes from 20m to 1m in about a boat's length.
Read more at http://www.ybw.com/forums/showthrea...d-you-like-your-GPS/page3#8tBfc50YTPTij9jU.99

It does in some places providing you go in with the sun over your shoulder.

FYI Boats who rely on GPS waypoints to enter the San Blas Islands will usually go aground at least 12 have done so in the last 3 years and most were total losses.

 

[mjcoon]

Isn't speed calculated from doppler shift of one of the frequencies?

http://gpsinformation.net/main/gpsspeed.htm

I've often wondered about this and have not seen an explanation that addresses my puzzlement. I understood that all satellites use the same frequency and are flying about much faster than we are on the ground/water. Hence doppler from one satellite would not give speed (velocity vector) and phase from multiple satellites gives a complex interference pattern.

Mike.

 

[lw395]

I've often wondered about this and have not seen an explanation that addresses my puzzlement. I understood that all satellites use the same frequency and are flying about much faster than we are on the ground/water. Hence doppler from one satellite would not give speed (velocity vector) and phase from multiple satellites gives a complex interference pattern.

Mike.

It's not simple stuff.
Basically, each satellite transmits a different code.
The GPS receiver has say 12 channels. Each selects only the code from one satellite and rejects all the others.
So you can extract a time difference and a frequency shift for each satellite.
Then find a solution which satisfies the relative speed and distance from several satellites.
Rejecting errors caused by electrical noise, interference, reflections etc.
Suppressing fixes which are miles from the previous fix.

Some years ago, I had a car with fairly early built in sat nav. That worked well because it used speed (and I think compass direction) from the car to aid the GPS. Whereas my TomTom in the same car would sometimes jump around when the signal was poor due to trees or tall buildings.

 

[mjcoon]

It's not simple stuff.
Basically, each satellite transmits a different code.
The GPS receiver has say 12 channels. Each selects only the code from one satellite and rejects all the others.
So you can extract a time difference and a frequency shift for each satellite.
Then find a solution which satisfies the relative speed and distance from several satellites.
Rejecting errors caused by electrical noise, interference, reflections etc.
Suppressing fixes which are miles from the previous fix.

That is not an "explanation" that satisfies someone who did electronics for his degree (and software thereafter). It just strings together the buzzwords used in many descriptions, like "channel" (much used in GPS advertising and I am sure is reflected in the GPS algorithms)...

 

[lw395]

That is not an "explanation" that satisfies someone who did electronics for his degree (and software thereafter). It just strings together the buzzwords used in many descriptions, like "channel" (much used in GPS advertising and I am sure is reflected in the GPS algorithms)...

Then Do Your Own Research.
The basic concept of code division multiplex allowing several data channels sharing a single carrier frequency is at the heart of it.
These are not 'buzzwords' they have actual meaning in the field.
Sounds like you did your degree a long time ago?

 

[mjcoon]

Then Do Your Own Research.
The basic concept of code division multiplex allowing several data channels sharing a single carrier frequency is at the heart of it.
These are not 'buzzwords' they have actual meaning in the field.
Sounds like you did your degree a long time ago?

That's very true ;-) And the concept of sharing a single carrier with multiple modulated data streams is OK too. (One could bandy about barely-understood buzz-words like "trellis coding", much used in North Wales.) And a single carrier could give a single doppler shift. But not a velocity vector. Just as cosmic red shift only gives speeds (+/-) towards Earth.

If I could have been bothered I'd have done my research when I bought my Garmin GPS-II+ (mentioned in the cited ancient web page) in 1997. Might try wikipedia...

Mike.

 

[lw395]

But the doppler shift varies from one satellite to another, and as any satellite moves around the sky.
That is how a single SARSAT satellite gets a fix on an EPIRB
I suspect ancient Garmins did not do speed by doppler.
In those days, receivers had to do a lot of averaging to get any accuracy. If you drove around a big roundabout at speed, the GPS might average your speed to near zero.

I'm told it's possible to get a doppler-based GPS speed reading which is consistently wrong, for instance on a windsurfer, pumping the sail at the right rate might cause the sampled speed to always be higher than the average speed.

Last time I fired up some UBlox GPS modules, the raw postion data coming out every second had a lot of random error. Tens of metres. There's a lot of processing goes on to give the display on a chartplotter or TomTom.

 

[doug748]

Good un, I understand that bit but that is as far as it goes - )

 

[Neil_Y]

50m? There's someone who's never sailed in Sweden.

be careful, not all the rocks are where they say they are. We raced through the rocks and aimed to give them at least 5m but ended up being much closer at times. I'll never know if it was the rocks or the GPS but there have been some boats parked on rocks a few times. http://www.stockholmarchipelagoraid.com/ our dagger boards draw close to 6' so we need as much water as many yachts.

 

[[70521]]

When you are looking for the west entrance to Cherbourg Harbour in less than 200m visibility as we were last Saturday, with no radar on board, Mark 1 eyeball alone isn’t enough. Knowing that the chartplotter is showing where you are accurate to a couple of boat lengths is very helpful. Only when we were actually in the marina could we see where we were. It was also helpful that as we knew where we were we could concentrate our efforts on looking out for where other boats might be. Fortunately nobody came within sight of us. The experience has persuaded the skipper that even if he can’t afford radar an AIS receiver is worth investing in. It certainly engenders respect for those that have gone before us without radar, ais or chartplotter.

The Mark 1 eyeball often doesn't spot ledges, or places where the depth goes from 20m to 1m in about a boat's length.

The above posts are interesting and demonstrate the move away from paper charts and the navigator's understanding the area they are sailing in.

In the old days we would have stood off Cherbourg until the fog lifted to give us enough to see or found a buoy and did the old trick of blind navigation and gone in very, very slowly from buoy to buoy. Is this dark art still practiced by most crews? Last time I was caught out in fog was between Guernsey and the shipping lanes to the north. Very open water and while we have an AIS transponder, nothing was anywhere close.

I suspect the moving sand banks of the many estuaries in the UK also have their challenges, I know going up and down the Exe can be fun for the first few times in the season as things will have changed over the winter.

 

[savageseadog]

The above posts are interesting and demonstrate the move away from paper charts and the navigator's understanding the area they are sailing in.

In the old days we would have stood off Cherbourg until the fog lifted to give us enough to see or found a buoy and did the old trick of blind navigation and gone in very, very slowly from buoy to buoy. Is this dark art still practiced by most crews? Last time I was caught out in fog was between Guernsey and the shipping lanes to the north. Very open water and while we have an AIS transponder, nothing was anywhere close.

I suspect the moving sand banks of the many estuaries in the UK also have their challenges, I know going up and down the Exe can be fun for the first few times in the season as things will have changed over the winter.

The old trick was to sail to a contour, a set depth on one side of a channel. If entering on normal side of the channel if the depth decreases, turn to port, if increasing turn to starboard.

 

[Birdseye]

Really depends on the accuracy of the charts/maps you are using at the time. Personally, within 50 meters is good enough for me.

The charts arent accurate to a foot or two. Indeed some locations ( Cornwall and N Scotland) had to be moved by half a mile of more when we went to WGS84 and had satellite piccies available. The places used for triangulation were out the further away you were from the origin.

The 95% confidence limit on the best accurate charts is +/- 5m say 65 foot. To that you need to add any GPS errors. See https://www.admiralty.co.uk/news/blogs/category-zones-of-confidence

 

[Uricanejack]

New long for fred rift

Isn't speed calced from doppler shift of one of the frequencies?



http://gpsinformation.net/main/gpsspeed.htm

Truthfully I have no clue how GPS works. Magic box, it just does. If I had to explain it for an examiner.
What the response would be?

The old transit system used doppler shift to derive a position from a single passing satellite. I believe this system is now defunct.
My limited understanding of the GPS system.
The Satellites are Geostationary. With fixed positions confirmed by fixed ground stations,

The GPS system is a hyperbolic navigation. System I think using time difference rather than phase difference. But for all I know both phase and time difference. Unlike the older hyperbolic systems GPS is 3 dimensional. So it can give altitude as well.
So the math involved is a bit more complex. But the magic box sorts this out. No CRT or Dials, special charts or rulers.
The different channels and No of satellites give lots of hyperbola, coarser and fine. allowing the magic box to eliminate wrong positions. resulting in very accurate position.

My personal deductions which might be bollox.
My magic box on my boat. doesn't need to use doppler shift. The Satellites are geostationary. So no dopplar shift. Unless of course they aren't geo stationary.

If my boat moves it would cause a small doppler shift to the signals received from the geostationary satellites. which if detected could be calculated but would still be subject to error cause by any other movements.

It would me mathematically much easier to compare successive positions and calculate my speed and course between calculated positions than trying to compute. Doppler shift.
There again I have no clue what I am talking about.