Another route to electronic timing gear?

[oipigface]

I was listening to a programme on Radio 4 the other day, which featured a recently developed system of surveying using GPS. I didn't catch its full name, but it’s an Ordnance Survey project, which uses GPS as its primary input. GPS is accurate only to the nearest couple of metres, but the journalist presenting the programme claimed that he had surveyed the boundary of a building site using a portable antenna and a laptop to an accuracy of +/-2cm! The increase in accuracy is achieved by the use of a kind of triangulation using signals from 115 ground stations scattered around Britain. Details were left undiscussed.

I can’t remember what show this was on, but it struck me that the accuracy claimed would be good enough for F3F timing gear. Big question is: How long does it take for the corrections to be calculated, transmitted and digested? The latency might be unacceptable.

[oipigface]

I’ve found the programme at last. It’s at https://www.bbc.co.uk/programmes/m000f6bp and the relevant part starts about 9:00 mins in.
The correction system is called OS Net. Can the Terminator be far behind?

[Brett82]

Hi John

It sounds interesting but I think the biggest problem we would face is getting a manufacturer of survey equipment to develop a custom unit we could use that would fit into out F3F planes. 

I do really like the idea of it though and a little while ago I was looking at this...  https://emlid.com/navio/

It would need a complete redesign so it wasn't sitting on top of and controlling a Raspberry Pi, and also to be a lot smaller. But a company like this might be more capable of developing something we could use. When you look at the accuracy of the Barometer for example, it looks like it would be accurate enough with low latency. They could even develop a centre base to work of a Raspberry Pi that would talk with the sensors and could even run a race management app, similar to Marks android app or the RiPi app being written by Pierre's friend. 

The only other issue I can see is linking the sensors to pilots and the centre base. Do pilots have their own sensors (which need to be compatible with each event locations F3F equipment, or do the event organisers have the sensors and pre-link them to a pilot and the base before the comp then handed out with the bibs. 

I love the idea of bringing in new technology though.

Brett

[oipigface]

I’ve given the matter a little more thought, Brett. I was thinking that maybe a transponder in the plane could be tracked by the GPS/OSNet. These will soon be required equipment in France. I don’t know if there are similar plans in Britain, and of course I don’t know much any of the more detailed issues that you raise. I don’t think transponders are very heavy or expensive. 
The only point I was really making was that tracking technology seems to have become much more accurate than it was a couple of years ago.

[Brett82]

45 minutes ago, oipigface said:

The only point I was really making was that tracking technology seems to have become much more accurate than it was a couple of years ago.

That for sure is the main point, how a unit is put together or how it works is kind of like saying how long is a piece of string. The way the tech is developing sure would indicate it wont be too long before we can integrate it into F3F.

17 minutes ago, pete beadle said:

Would it not be possible to have something like this that could be downloaded from the onboard GPS  of each plane at the end of each round perhaps?

Hi Pete

I think you missed the point of John's idea. The idea is to use a tracking system to integrate into the timing system of an F3F comp so you dont need people standing on a base pushing a button each time the planes passes a base. Not to download data later.

Rather, the GPS system would tell you when the plane had gone beyond 50 meters from the centre so you knew when to turn. That is why he was saying about latency being an issue, any delay in the GPS single would make it not accurate enough, also if it could only track to the nearest 5-10 meters it wouldn't be accurate enough.

It seems technology is advancing to a level where it could potentially be integrated in the near future. 

Brett

[wookman]

The latency and accuracy would have to improve a bit more I suspect before it would be better than the Mk 1 thumb.

On a slightly different tack, can you use rc telemetry gps with an alarm set at 50 meters distance for practice runs. I don’t know, I still fly DSM2!

[Brett82]

6 minutes ago, wookman said:

On a slightly different tack, can you use rc telemetry gps with an alarm set at 50 meters distance for practice runs. I don’t know, I still fly DSM2!

I think for practice it might be fine but you need to remember the fact the further out you fly from the slope the further you need to fly across it. Otherwise you could do some seriously tight figure of eights 50 meters out directly in front of you... not very accurate 

[wookman]

Absolutely Brett, that is why I was only advocating  it for practice.  I guess it would be possible to botch it a bit by saying fifty meters along the slope and ten meters out. Apply a bit school trig and bob’s your hypotenuse.  Having done all that it would only ever be a crude practice aid and would rely on flying a course a consistent distance from the hill. 

It’s a good job flying is bit like falling off logs otherwise next time I get the opportunity I would have forgotten how to!

[John Minchell]

Take a look at the RC Electronics GPS gear that Bernie does at E Soaring Gadgets.  The T3000 ground unit I have has a dedicated F3F task already loaded into the soft/firm ware.

Manual here https://www.rc-electronics.eu/wp-content/uploads/2019/02/RC_T3000_Manual_v1.2.pdf

See page 12 onwards, it may be good enough for you to use without re-inventing the wheel.  The airborn unit is the size of a couple of oxo cubes.

John M

[f3fisa]

Latency is always the problem with GPS. If you assume your plane is travelling at 25m/s then if using a 10Hz update rate on your GPS, which is very fast, your model would have travelled 2.5m.

Accurate GPS has been around for a long time in surveying etc., it's called differential GPS. Here is a link to explain

https://racelogic.support/01VBOX_Automotive/01General_Information/Knowledge_Base/How_Does_DGPS_(Differential_GPS)_Work%3F

As mentioned above in GPS Triangle Racing a GPS system is used to determine the course to fly, but even this system is a bit too slow at the moment to be of any use for F3f.

David

[oipigface]

Perhaps a bit of context would not come amiss.  There are several camera-based systems available that can replace a person operating a buzzer at a base when a plane flies past it. They mimic humans in spotting the plane from the ground each time it passes, then activating a signal to the pilot and to the central scoring computer. It must be about 30 months since I watched and listened to one that was set up in parallel with humans. The cameras always buzzed first, and there was no sign during about 60 flights of incorrect buzzing such as may be triggered by birds. 
There seems no point in trying a different route unless it is going to be at least as accurate as humans or cameras. It appears that cameras can be better than humans. So my question really is: Could a GPS-based system be better than cameras? In order for the FAI to be able to accept any automatic system they would need to be convinced that the course length was being accurately measured. Courses are routinely set up to be a little longer than 100m just to be sure, but in order to accommodate the kind of error that naked GPS is prone to, the course would have to be way over 100m - possibly as much as 106m. The OS NET correction system I think would reduce the likely error to an acceptable level. (1 or 2 cm as opposed to 2 or 3 m, according to the radio programme.)

57 minutes ago, John Minchell said:

The T3000 ground unit I have has a dedicated F3F task already loaded into the soft/firm ware.

That looks as if it might be useful as a practice device. Have you tried it at all? What’s the latency like? The accuracy is presumably standard GPS. i.e. not good enough for a competition.

33 minutes ago, f3fisa said:

Accurate GPS has been around for a long time in surveying etc., it's called differential GPS. Here is a link to explain

They didn’t call it DGPS on the radio, but that is what they were talking about. The impression I got was that it was quite new, but I can’t remember anyone from OS making that claim. It was certainly news to the journalist!

[simon_t]

Unfortunately the accuracy is only good for relatively slow moving GPS ‘objects’.  As Mr. Woods has pointed out, the GPSs available to us are pretty slow, and the algorithms and filters (Kalman) used to improve accuracy work against you when there are sudden direction changes.  One day maybe (after all there are GPS guided missiles which travel and manoeuvre very fast), but unless we get access to the best military GPS chips and algorithms  (unlikely) it seems some way away.

If model flying isn’t banished by regulation (looks like Portugal have just decreed that model flying will only be allowed from authorised, registered sites), then I look forward to future high speed GPS technology.  The best GPS triangle systems have an 18Hz update rate.

Simon

[wookman]

Anybody getting anywhere with a camera based system for practice or even competing?