adsH said:
...there can be an alternative to using a infrastructure that you described. using distributed computing power and placing inexpensive COTS Processors on the SAM sites and linking each Site with multiple links say fiber optics (even higher grade Cat Ethernet cables, you can always use compression) ( complexity of connections should justifiable).
That's sort of what I was postulating when I was referred to redundant processing and data pathways, and having each launcher control it's own missiles. However, I think that the amount of processing needed for actual targeting computations in a multistatic system would overwhelm inexpensive systems. If that's true, I think separate data processing centers would still more affordable for the heavy computing. But they would still be have to be redundant themselves, and linked with redundant pathways. So each launcher could handle it's portion of the task (receiving and processing the signals, forwarding the data, receiving target data, and controlling the missiles), but the bulk of the computing would be done in one of several data centers using server clusters. The datacenters could be buried underground in hardened shelters, and geographically spaced, which would make them more difficult targets.
There are pros and cons to each method- the servers can be software clustered and still be physically separated, but in increases the cost of each launcher/processor system, as each system would need to powerful enough to operate more or less independently, since the system has to be able to adjust to the data it's receiving. Possibly the overall network would degrade more gracefully than if there are say 5 large datacenters instead of 100 launcher/processor systems. But our infrastructure is getting more complicated, lol.
Another possibility would be to have multiple emitters that encode a time stamp into the signal itself (kind of like LORAN). Then the SAM could just take a GPS position, and calculate the geometry based on the known location of the emitters. Instead of geographically separating the receivers, you separate the emitters and work the problem in reverse. You would need a chain of emitters, and each emitter would work on a separate frequency to prevent crosstalk. This would make the cabling between sites unnecessary, and your launchers could be moved around as needed. Each SAM would be then autonomous and totally passive.
Or maybe a combination of both techniques- nodes of 3 receivers and 3 emitters working as an independent cell. Lol, my brain's starting to hurt...
The guy I mentioned earlier from the UW (John Sahr) showed that it can be done, at least in principle. He used a PC with a standard digitizer card and a single VHF transmitter. I have it on acrobat format, I went back and reviewed it, I should make a couple corrections. Timing was via GPS signals, as I mentioned, accurate to ~100ns (15m in range) and ~.01 Hz in Doppler (1cm/sec in velocity). The transmitter was in the Puget Sound basin, not on Mt. Ranier, so his receiver and transmitter was separated by a good sized mountain range (The Cascades). He demonstrated quasi-real time imagery of the ionosphere out to a range of ~1000 km, and routinely detected aircraft out to ~100km.
His system was a simple bistatic system, so the processing power was relatively low in comparison. But it does show that a high performance PCL system can be developed with inexpensive components. I think his entire receiver station was under $15,000. Of course, when you start adding multiple emitters/receivers, the algorithms get extremely complex, and I question whether inexpensive computers can really handle it.
adsH said:
...the reason you stated fibre optics was (i'm guessing here) Military applications require communication systems that are not susceptible to EM and other jamm interference. what i was trying to bring across here was that you can use simple cheaper COTS stuff and put together a sophisticated system. you can always use Extra shielded wiring. and of-course Fiber optics is the wiring of choice.
I assume the system is operating in a high ECM environment, which is why I go with FO. Also, because we are talking about distributing receivers over a large geographic area, at some point you need to use FO anyway. It's no more expensive than copper these days, so it is the obvious choice.
It's an interesting problem, anyway, fun to think about.
