Air Force’s ‘All-Seeing Eye’ Flops Vision Test

[ClearedHot]

Air Force’s ‘All-Seeing Eye’ Flops Vision Test

It’s the one of the most revolutionary — and one of the most chilling — weapons to come out of America’s decade of conflicts in Afghanistan and Iraq. Gorgon Stare, a new “all-seeing” camera system for aerial drones, is supposed to boost U.S. surveillance by an order of magnitude, by installing a hive of nine or more cameras under the wing of an Air Force Reaper drone. Gorgon Stare-equipped Reapers are meant to watch over a “city-size” area, while also simultaneously sending video feeds to dozens of “customers” on the ground.

There’s just one problem. Gorgon Stare doesn’t work as promised, at least according to the Air Force squadron whose job it is to test the new system.

In a draft report dated Dec. 30 and obtained by rogue military analyst Winslow Wheeler, the 53rd Wing at Eglin Air Force Base in Florida declared Gorgon Stare “not operationally effective” and “not operationally suitable.” Alleged problems include poor-quality video, glitches in the process for downloading video streams, and a small problem of the drone blinding itself with a laser.

This is bad. Real bad. The Air Force is counting on Gorgon Stare to help its squadrons in Afghanistan meet “insatiable” demand for overhead full-motion video.

Despite steadily adding drones — there are now more than 50 three-’bot “orbits” deployed in Iraq and Afghanistan — the supply is never adequate. “It’s like crack, and everyone wants more,” Army Brig. Gen. Kevin Mangum said of drone-supplied video.

Equipping a portion of its drone fleet with Gorgon Stare would be like adding hundreds of new drones, from the perspective of the soldier on the ground. Now, the Air Force might not get that boost.

A standard drone spycam takes a “soda straw” view of what’s beneath, focusing on a lone vehicle or a single home at a time. Gorgon Stare, on the other hand, uses a bundle of cameras, each one shooting at a very slow rate and at a slightly different angle. That allows the sensor to watch over a much larger area at once: about 36 square miles or so, according to some estimates.

“Gorgon Stare will be looking at a whole city,” Air Force intelligence chief Maj. Gen. James Poss recently told The Washington Post. “We can see everything.”

But that view might not be so clear. The 53rd Wing found 13 serious deficiencies in the system. To carry the Gorgon Stare pod plus the required processing pod, Reapers have to be stripped of other sensors and weapons and structurally reinforced. The layout of the underwing pod puts it in the way of the Reaper’s nose-mounted ranging laser, meaning Reaper remote pilots could accidentally “lase” and blind their own cameras.

Even when working correctly, the Gorgon Stare’s cameras are ”marginally sufficient to track vehicles” but “not sufficient to track dismounts [people],” the testers wrote. “In general, IR [infrared] imagery quality is poor, which yields marginal mission capability at night.”

Plus, soldiers on the ground could have a hard time capturing the Gorgon Stare’s video feed. Even if they do, a glitch in the system means imagery is “subject to gaps between stitching areas [where the camera images meet], which manifests itself as a large black triangle moving throughout the image.”

Gorgon Stare “cannot reliably find and track human targets; it has additional problems for moving targets, and the random location inaccuracy makes the system virtually unusable for prosecuting even stationary targets,” Wheeler summed up in an e-mail.

Gorgon Stare and other “wide-area airborne surveillance” (WAAS) systems came out of an Iraq-war imperative to track car bombs across an entire city. Spycams might not be able to prevent such attacks. But if a whole town could be surveilled at once, the car bombs could be traced back to their points of origin.

As the insurgency evolved, and the Afghanistan conflict heated up, the need for WAAS changed. Instead of spotting vehicles after the fact, the military wanted WAAS to find individual, dismounted attackers — in real time.

But with a resolution at least twice as bad as standard drone cameras — and a frame rate of just 2 per second, compared to 30 in standard-issue spycams — tracking individual militants might be too tough a mission for a WAAS system. “If somebody said Gorgon Stare could spot dismounts, they probably oversold it,” says a source familiar with the programs. “It was not designed to do that.”

The 53rd Wing recommended more testing and development before Gorgon Stare gets installed on Reapers in Afghanistan, something that is supposed to happen “this winter.”

We’ve asked the Air Force to respond to this report. We’ll let you know what they say.

Now, to be fair, tests are designed to uncover problems, not just rubber-stamp things that already work fine. And military testers are paid to be skeptics and trained to find even the tiniest glitch in new weapons.

That can result in unfair assessments of urgently needed weapons — and sometimes leads to clashes between testers and the broader Pentagon establishment. For example, testers for years declared the Navy’s EA-18G radar-jamming plane “not operationally suitable,” but the Pentagon pressed ahead with — and even expanded — plans to field the jet.

Similar assessments were made of the now-iconic Predator drone back in October 2001, when the testers found it to be not “operationally effective or suitable.” It wasn’t long before the robotic spy plane was making important contributions to the war in Afghanistan.

The 53rd Wing’s assessment of Gorgon Stare could be an example of over-stringent testers striving for unnecessarily lofty benchmarks. Or the Air Force testers are correct, and the Pentagon’s revolutionary all-seeing eye really is blurry and half-blind.

Update 7:11pm: “Gorgon Stare is in the first increment of a multi-increment program, and the second increment will increase the warfighter’s capabilities by range and resolution,” Air Force spokesman Lt. Col. Richard Johnson says in a statement, released moments ago.

The document leaked was a draft memo that was later revised in January.

The January memo includes three issues that we have identified and have fixes in place. The first was addressing critical Technical Order shortfalls; the second was Gorgon Stare Ground Station image and grid coordinate generation; and the third was Remote Video Terminal compatibility. We’re working all three issues and do not believe they will affect the deployment schedule.

Air Force leadership understands the importance of providing quick, timely and actionable ISR for the field. Gorgon Stare will not be fielded until the theater commander accepts it.

The Air Force takes its responsibility seriously because lives depend on the quality of the intelligence products that are produced.

[nsplayr]

Just read this on Danger Room & thought about posting. I think the expectations the company sold got a little blown out of proportions. At 2 fps how the F were you gonna track man-sized targets in an urban environment? Either way, the idea is potentially good but maybe the technology isn't quite there yet. Wonder how we can possibly digest all that video even if it does work...

[Skyclops]

Just read this on Danger Room & thought about posting. I think the expectations the company sold got a little blown out of proportions. At 2 fps how the F were you gonna track man-sized targets in an urban environment? Either way, the idea is potentially good but maybe the technology isn't quite there yet. Wonder how we can possibly digest all that video even if it does work...

Yeah, I wasn't following the rambling at the end of the article about testers being super meticulous and getting in the way of deploying it. 2 fps, inability to track man-size targets, etc. seem to be legit reasons not field it and they're doing the Pentagon a favor by doing their job. It would be useless and a waste of resources if the powers-that-be put this system out there as is.

Along with not being a mature technology, the article makes it sound like shoe-horning it into the Reaper, a plane that wasn't designed to carry it, isn't the best mate for the system. Obviously they're trying to use what they got. But putting a mature Gorgon Stare into some of the next-gen UAVs (or that Hindenburg carrying SkyNet) being tested out in CA has awesome ISR potential. Not a Global Hawk expert, but it seems like that would be a better platform. Even the MC-12.

[Guest]

"Air Force spokesman Lt. Col. Richard Johnson..."

He better have a damn good tactical. That name just screams possibilities.

As for Gorgon Stare...I'm sure it's nothing a few more billion thrown at the contractors won't solve. I'm guessing the Prime will somehow figure out a way to say the capabilites we need were not defined as threshold requirements but they can fix it for us as long as we cough up a lung and cut off a nut or two.

[tac airlifter]

Another technological marvel that fails to work as advertised. Suprise, suprise.

[Guest]

What does this button do?

Please note proper wear of the reflective belt.

and a small problem of the drone blinding itself with a laser

It's all fun and games until someone loses an eye.

Edited January 25, 2011 by Rainman A-10

[Skyclops]

Follow-up: problems have been addressed, but doesn't say which ones.

AF: Some Gorgon Stare Probs Fixed

The Air Force says in a statement that three of the seven problems cited in the testing memo have been fixed and that it will not field the system until theater commanders say they will accept it because Gorgon Stare will not be fielded until the theater commander accepts it. The service takes it responsibilities for developing the system very seriously, because, as the statement says, ” lives depend on the quality of the intelligence products that are produced.”

[HiFlyer]

No offense intended, guys, but this is mostly media BS from someone trying to make his name with the next big military boondoggle. The facts are a little different, as I know them. For instance, "To carry the Gorgon Stare pod plus the required processing pod, Reapers have to be stripped of other sensors and weapons..." Yeah...well, to carry its weapons, the AC-130 can't carry the C-130 cargo load either. Oh, wait...that's not its mission, is it? Well, the few GS Reaper sets that they're building have a different missiion than the strike Reapers. And by the way, their normal FMV sensor ball is still there, still being used; it hasn't been removed. Only some weapons on the wings are missing because the pods are installed instead. The fact is that this is a new system in its first OT&E, and they're finding a few problems, which they will probably fix fairly quickly. When was the last time we had a new capability go thru OT&E without some problems popping up? That's why we do OT&Es!

"This is bad. Real bad." Oh, come on now. It's one little set of pods, not the linchpin of American combat capability. No one likes to have a failed program (which this isn't likely to be in the end), but this dramatic hand-wringing is a little over the top. The same basic concept has proved very successful for the Army, with a similar set of capabilities called "Constant Hawk". GS combines that capability with expanded coverage, faster processing, and multiple near real-time comms. Perhaps there are a few problems to be worked out, but the ops concept is proven and sound.

Can't track dismounts? Well, it wasn't intended for that purpose as its primary mission. As they improve the sensor over time, I'm sure it will get better...all the other sensors I know of improved over time. Its called "spiral development!" The Phase 2 pods (yet to come) will be better because of what we learn with the first set.

This is yet another case of an overzelous author with little practical knowledge of a subject taking a report (and a draft report, at that) out of context.

Nothing here to see. Move along...

Edited January 25, 2011 by HiFlyer

[Guest]

Can't track dismounts? Well, it wasn't intended for that purpose as its primary mission.

So you're saying that wasn't a threshold requirement for IOC ?

As they improve the sensor over time, I'm sure it will get better...all the other sensors I know of improved over time. Its called "spiral development!" The Phase 2 pods (yet to come) will be better because of what we learn with the first set.

Were the phase 2 pod requirements in the original contract or are we paying prison sex for a pack of smokes style?

This is yet another case of an overzelous author with little practical knowledge of a subject taking a report (and a draft report, at that) out of context.

Nothing here to see. Move along...

I hope it is that simple, but i doubt it. There is something to see here if the pod isn't meeting threshold requirements AND/OR if we wrote an open ended contract that drives up the end costs far beyond the programmed budget meaning we ultimately get less than we need because we run out of money.

Money has always mattered but it really matters when the SecDef is looking for places to slash the budget.

[HiFlyer]

So you're saying that wasn't a threshold requirement for IOC ?

No, it wasn't. The requirement was to track small vehicle sized objects, although seeing (vice tracking a single guy in a crowd) dismounts were a nice to have capability if it could be done, and we may well get there yet. The "Increment 1 capability is just the "A-model" and improvements will be made as it is tested and fielded.

Were the phase 2 pod requirements in the original contract or are we paying prison sex for a pack of smokes style?

I'm not sure of the actual contract provisions, but the "Increment 2" development was intended to incorporate lessons learned from the first set, to the extent that technology can be applied and the cost is within reasonable boundries.

I hope it is that simple, but i doubt it. There is something to see here if the pod isn't meeting threshold requirements AND/OR if we wrote an open ended contract that drives up the end costs far beyond the programmed budget meaning we ultimately get less than we need because we run out of money.

Keep in mind that this isn't some extensive, long development effort via the standard DoD 5000 series procurement process. The AF asked the COCOM what they needed help with, they said they needed long duration area observation similar to the initial and widely applauded Army "Constant Hawk" program but with better comms and wider usability at the platoon level and needed it ASAP, and the AF Lab guys came up with something in about two years. It isn't perfect, its not without flaws, but if they can get it to function properly it will be an 80% solution that will help the COCOMs (BTW, it has a faster frame rate than the Army program, has real time comms to the grounds which the Army system doesn't have, it has IR which the Army doesn't have, and it has much more loiter than the Army system). There are dozens of similar programs like this developed on the spur of the moment that are already in use (called a QRC -Quick Reaction Capability). Some work well, some barely work, some don't work and are sent home and trashed. Most aren't even "Programs of Record" in the POM, and don't have formal requirement paperwork with "Threshold" and "Objective" requirement sets. That would add 5-10 years to the process and the need would probably disappear before it was fielded. These are QRCs intended to fix a COCOM CCs immediate priority need. The CENTCOM commander sends a letter to the Chairman, the staff work is done via a JUON (Joint Urgent Operational Need), and a solution is proposed to the Joint Staff, COCOM and Service; if the leadership agrees that it is a viable solution and the money is found (most of it through supplemental or reprogrammed funding), they try to implement it. Other similar programs include the MC-12 Project Liberty program, several IED detection and suppression devises, a number of unique sensor systems which help with specific target sets, etc.

That's what's wrong with the basic tenor of the article. The writer clearly doesn't understand what is going on. The original "whistleblower" is a member of a small "watchdog group" which tries (apparently) to expose waste in the military budget, and somehow latched onto the draft report of the first OT&E review and is trying to make it into a big scandle report. In fact, the OT&E is doing exactly what its supposed to do...find problems and fix them before fielding the system (or provide rationale to the SECDEF to cancel the development).

Money has always mattered but it really matters when the SecDef is looking for places to slash the budget.

Yep, and his staff has his eye on this one, too. But to develop stuff like this, you have to take some risks. This one may be worth it. Actually, the biggest problem may occur if it works too well; it will put out huge amounts of imagery data every day. The Vice Chairman recently spoke about the issue of looking at all that stuff. Too much for humans (it would take thousands of additional analysts for each CAP) and not good enough automated expoitation available. Of course, most of the initial use will be to send many small segments pumped to a Rover terminal in a small tactical unit's hands rather than long term intelligence analysis...tactical SA rather than intel analysis. Personally, I think that's fine...they're the ones on the ground who need to see behind the walls and around the bend! The advantage of this concept is that instead of dedicating one Predator to each unit, one or two such CAPs may be able to cover dozens of small units as long as they cary a ROVER unit with them...much more efficient that allocating a Predator CAP to ever patrol in a small area (Kabul, or the Bagram area, for instance).