Darth Vader
SENIOR MEMBER
- Joined
- Jun 19, 2011
- Messages
- 4,447
- Reaction score
- 3
- Country
- Location
The past decade has seen a phenomenal increase in the use of manned and remotely piloted aircraft (RPAs) using ever more sophisticated sensors and near real-time distribution capabilities. With armed RPAs now capable of providing quick-reaction air support in addition to providing a persistent intelligence surveillance and reconnaissance (ISR) activity in the US forces, combat ISR has grown by more than 3,000% in this period and in the last year alone, the USAF flew 350,000 hours on such missions. Since 2001, the USAF RPA fleet has amassed over one million flying hours and the service has said that it would now be quite unthinkable to conduct counter-insurgency operations anywhere without the situational awareness provided by a mix of ISR systems and RPA platforms. The operational ISR needs in Iraq and Afghanistan, and more recently Libya, have provided insatiable demands from ground commanders with growing complexities in attempting to deliver so much ISR data from a multiplicity of sensors and communications tools, while processing exploitation and dissemination (PED) in-theatre and globally has often approached saturation point. It is reported that 19 analysts are needed to fully handle the data from a single Predator mission feed but at any one time, scores of airborne ISR assets may be on station.
Some idea of the scale of these operations and the problems that have arisen in coping with the massive data and imagery flows might be grasped by considering the fact that the USAF already has 268 MQ-1 Predators and MQ-9 Reapers in service, with 48 new Reapers due to be delivered each year and the equivalent of 57 combat air patrols (CAPs) being flown daily. And these figures do not take into account all the other RPA and manned specialist sensor-carrying ISR aircraft contributing to this tidal wave of incoming data.
So what can be done to avoid information meltdown? The answer lies in the development of new artificial intelligence technology applied to even more capable sensors so that more processing can be carried out onboard the platform and within the imaging systems so that operators are only sent the information they want or need to see. This is a big call, however, and reliability (not to say infallibility) issues are paramount where mission outcomes are life critical.
First steps
The US Defense Advanced Research Projects Agency (DARPA) has been working with the Sierra Nevada Corp on developing and testing an advanced podded system called Gorgon Stare, carried initially aboard an MQ-9 Reaper RPA. This has nine digital cameras which are integrated onboard to provide total aerial coverage of an area the size of a city centre (with a 4km radius) with the ability to download 65 images simultaneously. The $15m programme is relatively low cost but has produced results that confirm that the technology is workable, based on tagging the images with a customised video system that can analyse trillions of bytes of video, reducing the analysis task from weeks to minutes. The video tagging involves the use of a full motion video asset management engine (FAME) with digital metadata tags appended to each frame of the video. This encodes the time, date and location co-ordinates and this can be layered with links to other data sources to give a high-speed smart search capability to narrow down the search task for human operators. Pop-up windows can display different images to identify contents and the history of movements can be called up to see patterns of life. Much of this can be fully automated up to the point where the system will deliver just what the operator needs. It has been reported that in the evaluation trials carried out at Elgin AFB with the USAFs 53rd Wing, there were some teething problems resulting from combining multiple video images from multiple cameras, as well as night vision issues, real-time slippages and compatibility problems with hand-held Rover ground terminals but, with such highly integrated systems, problems such as these might well be expected and it is believed that all key issues are resolved. The first Gorgon Stare equipped MQ-9 is due to have entered operational service around the time this article is published, with two more due to be delivered.
All in the mind?
Alongside this development another DARPA funded project, Minds Eye, has been developed to automate the sensors ability to recognise images and to make comparisons with previous images to understand and identify movements and patterns. This is within artificial intelligence territory and is crucial in the steps towards fielding systems that can take a wide-angle look at what is happening and make note of how people and things move about over a period of time. At present most of these operational requirements are met by skilled specialist operators filtering out unwanted data by combining experience and intuition with the visual inputs from the Mk1 eyeball and increasingly capable computer assisted sensor displays. But human limitations demand far greater assistance from automated means if vital data is not to be missed or data flows become too massive to cope. There is no doubt that this trend is heading towards the robotic battlespace environment and at a far faster rate than many realise.
Gorgon Stare is just the first of a new generation of intelligent ISR sensor systems within what is known as wide area airborne surveillance (WAAS) that is set to change the pace and nature of automated data exchange above the front line. The US Army is about to field three rotary wing UAVs (Boeing A160 Hummingbird platforms) carrying the autonomous real-time ground ubiquitous surveillance imaging system (ARGUS). Each aircraft will carry a 1.8 gigapixel camera package with 92 simultaneously recorded video images covering a 36 square mile surface footprint. The system can collect and absorb over 75 years worth of HD video every day! (six petabytes of video) The Hummingbird will be able to loiter at an altitude of around 15,000ft for up to 20 hours, so a trio of platforms can provide continuous coverage.
A third new super-intelligent ISR system is MAV 6s Blue Devil, a $211m 350ft long airship that will be able to stay aloft for a week at a time carrying 96 cameras and can generate 274 terabytes of data every hour. The sensors will all be integrated with an onboard super-computer that can process up to 300 terabytes an hour and distribute processed data within 15 seconds. These levels of data are truly vast so only by auto-selection and artificial intelligence can useful levels of appropriate information be handled in packages that will enable operators to retain ultimate control over essential decision-making actions. No doubt new command and control challenges are just around the corner, as the increasingly robotic direction of system development will require very sensitive handling if combat ISR missions are to be entrusted more and more to machines. Science fiction horror films such as i-Robot may not be so far fetched after all.
Re-think on RPAs
The US Services are re-examining planned RPA programmes as a result of financial pressures on the defence budget and in the context of new enabling ISR technologies, as well as likely changes in the nature of future conflict. With the diminishing prospect of political support for operations such as in Iraq and Afghanistan likely to forthcoming in the foreseeable future, the DoD and individual US services realise that todays accepted superpower domination of air space in a threat-free environment cannot be relied upon to continue indefinitely. As a result, the present $20.4bn budget proposals for up to 400 Reapers and 55 Global Hawks by 2020 will need revision. Even back in 2009, the then Head of USAF ISR, Lt Gen David Deptula, said that the RPA master-plan would have to roll out to 2047 offering a System of Capabilities with a roadmap developing payloads, networks and processing tools with more modular systems, standard interfaces and more automated systems to reduce the human footprint. Gorgon Stare, Argus, Minds Eye and Blue Devil all point in this direction. It is thought widely in DoD circles that the any Reaper replacement will now have to be both jet-powered and stealthy to be survivable in the contested air space of the future. Trying to get more capability for no more cost is difficult but fewer platforms and more intelligent systems is a possible way forward. The added costs of supporting stealth in an RPA puts ISR decision-makers in an awkward corner, as they fear it may well be needed but know it could push development costs into new procurement danger zones. The Reaper family has proved to be excellent value to date but its slow transit speed reduces its ability to re-deploy to another location quickly in theatre and it would be very vulnerable to any half-serious air opposition. But it is not just the platform configuration that is key to the future. This is to be found in breaking through the bandwidth limitations on data distribution by more onboard processing and image and data selection.
The era of ISR stovepipe solutions is coming to a close and PEDs will move to centre-stage. The USAF believes that seeing and tagging by automated systems will allow more opened up system architectures to make it all doable. There will always be too much total data available to use immediately but, if it is recorded, accessible and has been catalogued it can be called up quickly when needed, or presented automatically as information that should be noted. There really is no other way forward for unless the pressure on military distribution networks and their human operators is eased through smarter onboard technological solutions the utility of ISR will be fatally compromised.
