http://www.au.af.mil/au/afri/aspj/di...df?source=GovD

 

Yeah, I hope he got a A on his paper . But it is funny how an RPA driver is telling us we don't need RPA drivers.

 

No, not really. As a matter of fact, it actually makes sense since one who has accepted an advancement of technology into a specific role is more likely to accept the continued advancement of that technology.

For a pilot to advocate against a specific technology simply and solely because it may eliminate his position is no better than those calvary officers who based their arguments that the horse mounted soldier would never disappear on the fact that they were calvary officers. I, for one, think that we are a LOOONNNGGG way from having machines that I would trust with a shoot/don't shoot decision (too many chances of anomalies with synthetic VID, EID, etc), but I would rather debate the subject based on intellectual honesty and an open mind than on parochialism.

 

Holy frat potential!

 

It will be a long time coming...for autonomous weapons release and PAX transport. ISR and cargo-only we'll see sooner. No military leader in his right mind will commit vast resources to a force structure which would be instantly grounded by political repercussions the first time it "decided" to shoot down an airliner or "auto-bombed" a pre-school.

If we were still in the cold war, you could make a case for an selectable autonomous mode where you "take the leash off" the automated weapons when the time is right...such as an open-ocean naval battle far away from collateral targets. Maybe we could buy some to fend off the PRC in the westpac. But in modern littoral and low-intensity urban conflicts?

In theory an autonomous air-to-air mode *could* be more effective without the man on board...saves a little weight, no g-force limits, and in the relatively narrow tactical realm of A2A a computer could weigh all of the available options and pick one faster than a man. But that's theory and assumes a bunch of things...

- No design errors
- Bad guy can't find a loophole in the decision tree
- Bad guy can't hack the system
- Target ID can be performed accurately
- No (or limited autonomous) multi-role: RELIABLE ground attack capability in a dynamic/CAS role is far beyond the realm of any plausible artificial intelligence systems in the works today. Unless you're talking about a bomb-truck mode with a guy on the ground calling the shots, which isn't really autonomous anyway.

We still haven't come up with anything that beats the flexibility of the human mind.

 

+1 above.

Anyone who has done computer-based training in the innocuous environment of one's own airline training center has had the system crash, while doing 1-g, and zero knots.

Everyone has had an FMC freeze-up at least once, some have had a failure.

Electronic Identification (EID) relies on radio signals of one kind or another, whether that be radar or interrogation of secure and non-secure squawks.

And radio signals (with the exception of HF, and we know how difficult it it is to use THAT) like line-of-sight. Even with LOS, no guarantees. Ever been in your home town, a regular location near a cell tower, and can't get your cell phone to pick up a signal?

Point being, the autonymous killers of Terminator will work just fine....in a world without mechanical or electrical failure or disruptions, or the limitations of physical law.

And that is about as likely as an honest and principled politician or car-salesman.

 

Nice Weapons School paper. I'll stack it next to the ones I still need to read written by Space WICS grads.