F-35 Lightning II: Busting Myths video

Discuss the F-35 Lightning II
  • Author
  • Message
Offline
User avatar

popcorn

Elite 5K

Elite 5K

  • Posts: 7703
  • Joined: 24 Sep 2008, 08:55

Unread post29 Dec 2016, 23:09

MADL employs daisy-chaining transmitting at much higher bandwidths IIRC from previous discussions.
"When a fifth-generation fighter meets a fourth-generation fighter—the [latter] dies,”
CSAF Gen. Mark Welsh
Offline

hornetfinn

Elite 2K

Elite 2K

  • Posts: 2814
  • Joined: 13 Mar 2013, 08:31
  • Location: Finland

Unread post30 Dec 2016, 09:48

popcorn wrote:MADL employs daisy-chaining transmitting at much higher bandwidths IIRC from previous discussions.


Yes it does. Link 16 for example is TDMA-based data link working in L-band.TDMA is channel access method where link nodes (like individual fighters) share the network medium. Basically each node has time slots where they can transmit their data and all the other time they are receiving only. So it has relatively high latencies and is quite slow data link. Standard Link 16 has max speed of about 1 Mbps. It might transmit data once or twice every second. This is good enough to share for example track data for a large number of targets between fighters and also to and from C2 nodes like AWACS.

MADL is directional data link using K-band. This means the bandwidth is far higher than in Link 16 and the nodes do not share the medium but they send data directly to several other nodes simultaneously. That data is then daisy-chained to other nodes almost immediately. Daisy-chaining is given because it uses directional beams and thus uses basically mesh network topology (like point-to-point topology with multiple nodes). This means the bandwidth is likely hundreds of times higher (depends on exact design of the system) and latencies are very low. Other K-band data links usually have several hundred Mbps maximum speed. Latencies are almost non-existent in this kind of data links as they do not have to wait for their own time slot to transmit.

A good way to think about this is that Link 16 is like 2G cellular network. MADL is likely something more ike 4G LTE Advanced in comparison. 2G is good for many basic tasks (like phone calls or SMS) but you can do so much new things with 4G LTE-A that are impossible or at least unpractical with 2G devices.

A good quick reference to tactical data links, including 5th gen ones: http://www.northropgrumman.com/Capabili ... orking.pdf
Offline

hornetfinn

Elite 2K

Elite 2K

  • Posts: 2814
  • Joined: 13 Mar 2013, 08:31
  • Location: Finland

Unread post30 Dec 2016, 12:22

gideonic wrote:This is just a speculation, but, I would be very surprised if polling the other nodes for data is the usual approach. Considering the speed, at which these aircraft fly and the extra latency this "round-trip" would add, the data would become outdated very quickly (at least for air-to-air targets). I'm not saying, it wouldn't be used at all, but I doubt it's the most prevalent one.

Also, in 4-ship sensor fusion, there are 12 separate datalinks (each aircraft connected to 3 others). Can you imagine the overhead of each of the aircraft constantly polling all the other ones? Such "micromanagement" of the other nodes seems unnecessary, badly scalable (to future 8-fusion and whatnot) and kinda a violation of some coding principles (TellDontAsk for instance). In some cases it might also eat noticeable bandwidth in it's own right (though latency is surely a much bigger concern).

The aircraft in a 4-ship, have the same hardware and software, so most of the time they should already know, what the other nodes would find interesting, in their data-stream ... and if they don't send anything, the other nodes can presume, it's because they have nothing of interest to send, no need to ask for it.

And there is more ways to go about it, than only sending confirmed tracks, or entire raw data as you mentioned. Probably they have something akin to a prioritized message-queue, where they send actual tracks and confirmed targets first, then (if current link quality and bandwidth allows) whatever additional metadata about those tracks and lastly other (already somewhat filtered out) "points of interest", that you describe. Such an approach would, have less latency, and still saturate the bandwidth as much as it could.

This is just a simplified example of course. There is no way to know how, they would actually prioritize the data. Regardless, having tons of data, and not enough bandwidth to share it (or processing power to filter it all) is a very common problem in the Software Development world. If it also needs to be realtime, then stream-processing is the only game in town pretty much.


I agree with this. Going backwards from tracks there are radar plots (hits) which are radar returns that exceed set detection threshold from which radar tracker generates and maintains tracks. Plots have just bearing, range and also processed doppler information. Tracker combines a number of plots to a track and then additional plots update the track data if they conform to existing tracks. These plots could probably pretty easily be transmitted over MADL network to give additional information. It could be used for example to direct APG-81 or EOTS to scan for targets in certain direction or just give sensor fusion more data to fuse which could improve weak target tracking or overcome powerful jammers.

The whole radar signal processing starts from raw radar video which contains all the signals radar receives. This is then processed to filter out unwanted noise and clutter and given to plot extractor to find return signals that exceed set thresholds to find potential targets. Sharing the radar video over MADL even in processed form sounds unnecessary and rather wasteful as the amount of data would be staggering and likely benefit would be low.

I think MADL likely exchanges tracks and plots (along with all kinds of other informatioN) but due to low latency and high bandwidth, they are much more useful to sensor fusion than with legacy high-latency and low-throughput (comparatively) data links. Legacy data links do not usually exchange plot data as they'd eat too much bandwidth and because latency would make them unusable for anything really. Tracks and plots might also contain things like doppler data which might be interesting for sensor fusion.
Offline
User avatar

popcorn

Elite 5K

Elite 5K

  • Posts: 7703
  • Joined: 24 Sep 2008, 08:55

Unread post05 Jul 2017, 22:23

Hey Dragon... what about a video centered around the recent Paris Air Show demo just to rub it in? :mrgreen:
"When a fifth-generation fighter meets a fourth-generation fighter—the [latter] dies,”
CSAF Gen. Mark Welsh
Offline

magitsu

Senior member

Senior member

  • Posts: 417
  • Joined: 12 Jun 2015, 22:12

Unread post05 Jul 2017, 22:41

hornetfinn wrote:A good way to think about this is that Link 16 is like 2G cellular network. MADL is likely something more ike 4G LTE Advanced in comparison. 2G is good for many basic tasks (like phone calls or SMS) but you can do so much new things with 4G LTE-A that are impossible or at least unpractical with 2G devices.

Hmm... I remember reading that for MIDS-JTRS, about to be installed to Finnish Hornets, max throughput is around 1MB.
Maybe it's much too simple to consider it "broadband of the skies" based on this.
Previous

Return to General F-35 Forum

Who is online

Users browsing this forum: No registered users and 5 guests