Show Posts

You can view here all posts made by this member. Note that you can only see posts made in areas to which you currently have access.


Messages - Pidea

Pages: [1]
1
OGN-R PilotAware / Re: Map of PilotAware / 360Radar ground stations -
« on: July 20, 2020, 11:44:55 am »
Didn't we email back and forth about this ?  Pretty sure that I sent you some maps etc having worked out the likely coverage based on your new proposed receivers plus the existing ones in the area.

2
OGN-R PilotAware / Re: Map of PilotAware / 360Radar ground stations -
« on: July 20, 2020, 10:26:46 am »
There isn't a specific map for European receivers as yet.  Although the locations shown on the map above are correct (give or take a deliberate error for privacy) so you know the approximate locations.  From memory, all of the European ones on the map are helping to MLAT aircraft.

To successfully MLAT an aircraft we need it to be seen by at least 4 receivers at the same time.  In flat areas this is easier than in hilly areas.  The height of the aircraft also makes a difference.  The higher it is, the more chance of it being seen by multiple receivers.

3
OGN-R PilotAware / Re: Map of PilotAware / 360Radar ground stations -
« on: July 20, 2020, 09:00:36 am »
The map located here https://radar.lowflyingwales.co.uk/pilotaware-contributors/ is an internal tool that we use to keep an eye on the PilotAware groundstations that connect to 360Radar to either contribute to our MLAT coverage or pull from it.  It was primarily put together to show where it was likely that an aircraft would pick up transmissions from a PAW groundstation.  It was based on code used to create other maps and as a result also happens to show the status of any receiver whose name starts with PW.

On the map, blue pins are PAW ground stations contributing to MLAT positions because they have a second 1090 MHz SDR whilst red ones are either offline OR are simply pulling MLAT positions and are not contributing to them OR are sending data to our European server.

Receivers that show as red can do so for several reasons - often users will name a station whilst setting it up and then will rename it again.  A good example is PWPWThame which was correctly renamed to PWThame in November 2019.  PWThame has uploaded recently but was last seen yesterday (July 19th) as indicated by the Last Seen time.  PWPWThame last uploaded, according to the Last Seen time, back in November 2019 presumably when it was correctly renamed.  With nearly 1,200 360Radar contributors there's a fair amount of churn where users install a system and then reinstall it at some point later choosing a different name or spelling it differently.

In an ideal world the system would track these automatically but it doesn't and I then end up having to manually investigate why a station appears to be offline.  I am currently in the process of cleaning the database of receiver names so that an automated email alert can be sent when a receiver hasn't uploaded for a period of time.  This will contain the receiver name, latitude, longitude and altitude so that should the user need to re-install they can use the same values.

The Last Seen time is based on when the receiver last contributed to an MLAT calculation and to work that out we need access to the relevant server log files.  Since this is an internal tool that was quickly thrown together for a specific purpose it doesn't access the data from our European server which is why all European stations are showing in red so in the case of European PAW ground stations this map can be ignored.

The Last Seen time will vary according to the receiver location so those receivers in more remote areas where there is less Mode S traffic may have Last Seen time several hours old.  Since the demise of regional airline FlyBe (who had quite a large Mode S equipped fleet) the 360Radar network spends most of its time tracking mostly GA and military Mode S aircraft so once GA aircraft stop flying most groundstations will not contribute to an MLAT position until either a military aircraft flies over or another Mode S aircraft comes into view.

So does that mean that this map is useless ? No but it could be better.

The map suggests that PWThame last uploaded yesterday morning (July 19th at 09:22) indicating that there probably is a problem with that receiver as there were plenty of Mode S aircraft around to MLAT yesterday but with lots of dead receivers on the page it's misleading which is why I'm working on an email alerting system to notify that PAW groundstation owners that their receiver is down.

4
OGN-R PilotAware / Re: 360 Radar contribution
« on: January 13, 2020, 07:28:08 pm »
You rang m'lud ?

Phil

5
General Discussion / Re: 360radar network
« on: June 05, 2019, 08:54:17 am »
Another friend with Mode S and ADS-B Out always shows significantly varying ground speeds when monitored with FR24. 0.0 groundspeed in 360radar.
Just trying to understand the differences in ground speeds seen and equipment etc.

Tony, we've talked about this via email.  When FR24 store your tracks they store as little information as possible so store the timestamp, lat/lon and altitude only.  When replaying a route they will calculate the speed, track and VSI between these points.  They are not the actual values sent by the aircraft.  If these points were every second then the calculated values would be accurate but they aren't.  They compress the data in order reduce its size and down sample by moving the points to where significant changes occured e.g. top of climb, start of descent, significant change of heading etc so the FR24 values that you see are the averages between those points.

The 0.0 values that you refer to regarding 360Radar are the from a rudimentary (and incomplete) logging system that is still work in progress.  They do not reflect the values that are produced by the MLAT server which are, in any event, of no interest to PilotAware anyway.

FYI, the L4 Cub was visible to mid channel (flying at 3000ft) but then disappeared completely (he did make it safely to Cherbourg!).
Can't be many 360radar receiving sites in northern France as a few would be needed to MLAT his Mode S only transmissions.
Tony

We have a handful but not enough not MLAT down to 3,000 ft.  360Radar is focused on developing the most comprehensive network of receivers in the UK and that's something that we've achieved.  From what I can determine we have more than FlightAware now, more than PlanePlotter and probably more than FlightRadar24 although it's hard to know as they don't publish their receiver locations.  We will slowly extend coverage into Europe this year and next.

6
General Discussion / Re: 360radar network
« on: June 05, 2019, 08:41:38 am »
A screenshot I took the other week from 360Radar MLAT track of a Partenavia over the Harlow area clearly doing something like Google earth surveying. A precision pattern is clearly being flown by the aircraft

In this case where tracks can be referenced against each other, it's quite clear to me that the MLAT data is pretty damned (sufficiently) accurate!

And just to point out that screenshot was taken not from 360Radar but from the local VRS using the raw MLAT data.

7
So what's on 8080, 8081 and 80 then ?

8
If you've got this on the internet and you've got ports 80 / 8080 and 8081 open then it's highly likely that your traffic is being siphoned off.  It's really easy to build a Google search query to find VRS instances online and then pull their data without the owners knowing.

VRS is great when running on a closed network with a small handful of users but it's very inefficient when run on the internet.  For every single connection to VRS it spawns a new thread.  Even on a decent server it'll start to slow down when it gets to between 50 and 100 simultaneous users.  That's why 360Radar runs through load balancers to spread the load across multiple servers and why I've also removed a lot of the duties from VRS such as the image generation as that is extraordinarily inefficient.

If you must run it online then password protect your feed to stop people leeching your data.

9
General Discussion / Re: 360radar network
« on: June 04, 2019, 01:38:45 pm »
Quote from: mariko
Understood. So you can multilatering the horizontal position of an A unique squawk code, but without vertical position. Or you have a different way to combine A answers with C answer... I'm getting curious...  ;D

You'll have to stay curious I'm afraid  :)

10
General Discussion / Re: 360radar network
« on: June 04, 2019, 12:53:23 pm »
really interesting! The tests consist to assign a C answer to an unique A answer with the same signal strenght, and use MLAT to determine horizontal position (as for S trasponder)?

Nice idea but unfortunately there's no easy way to get the signal strength from the receiver for a Mode A/C reply

11
General Discussion / Re: 360radar network
« on: June 04, 2019, 12:31:39 pm »
How!?!? Is it Really possibile with a omnidirectional aerial?

Yes, very early tests suggest that it would be possible in time for some aircraft as long as they use unique squawk codes.

12
General Discussion / Re: 360radar network
« on: June 04, 2019, 11:11:53 am »
Thanks to Lee for pointing me at this.  I'm Phil, the guy behind 360Radar.

360Radar consists of a network of approximately 900 Raspberry Pi-based receivers and around 40 Windows-based ones scattered across the UK from the Outer Hebredies, Orkneys to Lands End all the way to Dover.

Each of these runs our multilateration client and they collaborate together to calculate in realtime the position of Mode S aircraft (those ones that don't transmit their locations).  In order to successfully MLAT an aircraft we need a minimum of four receivers to see the aircraft at the same time.  Despite this challenge we can MLAT down to less 1,000 ft in some parts of the country but this depends on the terrain and the number of receivers (and we're always looking for more locations to site a receiver especially if you're in a remote location).

An increasing number of the PilotAware groundstations also run our MLAT-Client (they need a second SDR fitting) which helps to improve our coverage as many of these are on airfields and these tend not to have too many obstructions around them.  As a result these provide some excellent coverage.

Some of our receivers also upload ADSB messages from aircraft that do transmit their locations.  We don't need as many of these as only one receiver is needed to locate these aircraft.

We also get a feed from the OGN network which allows us to track PilotAware equipped aircraft, gliders and any other FLARM equipped aircraft such as the RAF Grob Tutors and Tucanos.  No 360Radar receivers (unless a PilotAware one) receive PilotAware data as they would require a second SDR and aerial.

So what happens to all this data ?  A subset (traffic below 10,000 ft) of this is then sent to PilotAware for rebroadcast by PilotAware groundstations whilst the full feed is sent to 360Radar and merged with all the other data for our contributors and subscribers to view.

But what about those aircraft that are only Mode A or Mode C ... well, we're working on tracking those too ;-)


Pages: [1]