The Mantis can fly for 24 hours without refuelling, do the surveillance job of four helicopters, acquire its own enemy targets and deliver a deadly payload - all without a pilot and crew. But should we be afraid of Britain's new robotic air force?
The Mantis carries no human crew. The plane is controlled by a set of computer components not that far removed from the chips and boards inside a high-end personal laptop
The aircraft is the size of a medium range bomber, with huge grey wings stretching 70ft across the hangar. It looks for all the world like any conventional aircraft - the wings, the nose, the wheels are all familiar. The engineers standing in front of it are dwarfed by its bulk. Modules beneath the wings can carry air-to-ground missiles and precision-guided bombs.
Other racks on the nose can carry surveillance equipment so advanced it can decrypt and listen to mobile phone messages instantly as it flies over, at heights of up to 60,000ft. It takes a while for you to notice the most important fact - there is no cockpit. There are no windows anywhere on the craft, - and no doors.
The Mantis carries no human crew - one of the reasons it can stay airborne for 24 hours. The plane is controlled by a set of computer components not that far removed from the chips and boards inside a high-end personal laptop. But unlike the American Predator and Reaper drones now flying over Afghanistan and Pakistan, this isn't flown by pilots via satellite control from a bunker outside Las Vegas. It flies itself.
The aircraft is sitting in the hangars of BAE Systems, just outside Preston - next to an airfield where Eurofighters are shooting vertically upwards from a take-off strip. The site is vast, with limousines ferrying suited executives from one part to another, and visitors carefully shepherded only into the areas they are cleared to see.
To enter Mantis's hangar, you have to pass through a glass cubicle that scans for any transmitting equipment - phones and cameras are strictly forbidden. A recording suddenly blares, 'Mobile phone detected!' as one of my hosts remembers he has a BlackBerry in his coat. I'm allowed to see Mantis, but not to know where the aircraft is currently flying.
The Mantis on the runway
Mantis isn't a 'drone'. It's a robotic aircraft. It's among the first of a new breed of armed UAV (unmanned aerial vehicle) that can take off , fly, plot courses and even acquire targets for itself, and the UK is at the forefront of this new technology. The Mantis only needs human beings for one thing - to pull the trigger.
The wars in Iraq and Afghanistan - with their political pressure for low casualties - have caused an explosion in demand for craft that need less and less human input. The spectacle of captured pilots was a staple of the first Iraq war and other conflicts around the Middle East. It's a vision that has been absent from the news this time round for one simple reason: there are now fewer pilots.
Robotics is a revolutionary technology on a par with gunpowder and the atomic bomb. It's another genie you couldn't put back in the bottle
Laptop parts, satellite connections and software are doing it instead. There are now 12,000 UAVs, used for everything from surveillance to search and destroy missions. In less than a decade, the business of unmanned aircraft has gone from being a minor, specialist sector to being worth £9 billion.
In the training missions that BAE is allowed to discuss with me, Mantis takes off entirely independent of its crew. When airborne, it is controlled either from a base in the UK or from a command-and-control centre so tiny that it fits inside a packing crate, which can be flown to a combat theatre inside a transport aircraft, with a commander and crew ready to deploy.
A satellite relays information to the Mantis, while pictures, video feeds, infrared images and decrypted phone calls come back from the battlefield. Six screens back on the ground off er a Mantis-eye view, a map and a set of geometric patterns showing the Mantis's orders.
Identifying potential targets
Previous generations of surveillance craft deluge intelligence staff with so many pictures that up to 160 back-room staff are required for each aircraft, but Mantis decides for itself what is interesting. A single Mantis can do the surveillance job of three or four helicopters or three Nimrod jet crews.
While it's in flight, no one controls Mantis with a joystick. Details of the mission are copied on to a memory stick and loaded into the control system's computers by the commander. In training two Mantis operatives can oversee up to three aircraft at once.
A video of BAE's software in action shows the aircraft targeting a line of trucks from miles above the Australian outback, with squares appearing over vehicles showing that they are objects of interest while Mantis flies over to investigate. The software inside Mantis has decided that they are moving, that they are in an area they shouldn't be and that they match its criteria for further investigation.
For a terrorist, or a lone psychopath, the idea of a vehicle that could launch, find targets and attack autonomously must seem like the ultimate risk-free weapon - a suicide bomb without a suicide bomber
Until now, the British Army has relied on American and Israeli drones, but Mantis is home-grown technology. In just four years the Mantis family of aircraft has gone from laptop components strapped to a second-hand glider bought in Wolverhampton to an operational spy plane due to enter full service in 2015.
The process has cost £124 million, and development has been spread across a team of British companies, including Rolls-Royce and QinetiQ, and British universities, such as Loughborough. At least two Mantis planes are being tested in the air right now over combat zones, although BAE is not allowed to say where. Other drone companies such as Boeing, Northrop Grumman and Lockheed Martin are making their own autonomous versions - but none can match demand.
First flight of the Global Hawk RQ-4
Autonomous machines save money, save pilots' lives and point to a future where stealth-enabled unmanned fighters and ultra-long-endurance surveillance planes can almost remove human beings from the aerial battlefield. But this technology has largely appeared without governments or the public questioning it. Can a chip make split-second decisions as well as a highly trained pilot? What happens when these systems fail? And worst of all - what happens when one falls into the wrong hands?
Unmanned aircraft have been used routinely since World War II, when the Germans used a remotely piloted bomb drone known as the 'Fritz'. But the market has exploded in the past ten years. There are 43 nations currently developing their own unmanned vehicles, including China, Iran and Russia. Some predict-that the market will hit a value of £53 billion - and the U.S. Army already predicts that its air force will be 80 per cent robotic by 2020.
Although drones are widely used, air forces tend to be nervous about letting them fly under their own steam - so highly trained pilots are still used, with a full back-up staff to ensure that nothing goes wrong.
'The way the U.S. military likes to do things, current attack drones require up to three pilots to operate - fastjet combat pilots, who are rare and expensive front-line assets,' says Steve Worsnip of BAE Systems. 'But the RAF doesn't have the luxury of those sort of numbers. They simply can't fight wars that way.'
The ground crew track a drone's flight path
'The human role isn't disappearing, but it is changing,' says PW Singer, a former Pentagon weapons adviser and author of Wired For War: The Robotics Revolution And Conflict In The Twenty First Century, 'Humans are no longer making decisions in the here and now; rather, they are simply supervising.'
In the U.S., more drone pilots are now being trained than actual pilots - and the degree of autonomy exhibited by the aircraft has increased to the point that new controllers don't even need to know how to fly. Many are videogamers, more than ready for the dehumanised, computer-assisted world of drones.
'People have an inherent fear of autonomous aircrafts, but as we face technical and battlefield problems, the solution is more and more autonomy,' says Singer.
'One of the problems right now is that unmanned systems such as Predator are gathering enormous amounts of data. We're about to add "Gorgon Stare" to Predators - an array of 12 video feeds. We can't keep up, but give the sensors more autonomy and they will decide what they send. A lot of the scientists told me that robotics is a revolutionary technology on a par with gunpowder and the atomic bomb. It's another genie you couldn't put back in the bottle.'
The BAE Mantis isn't the only unmanned aircraft that can operate independently. Global Hawk RQ-4, made by Northrop Grumman, is a huge, high-altitude craft that has been flying over Afghanistan for a decade, and has no need for pilots, either in the air or on the ground. It was the first unmanned vehicle capable of flying itself, and has completed more than 30,000 combat hours overseas. Its makers seem off ended by the use of the word 'drone' and refer to it as a 'robotic aircraft'.
Twelve years ago, a prototype of the RQ-4 Global Hawk was flying at 60,000ft above the Atlantic Ocean, near the east coast of the Azores. Its flying altitude is almost double the ceiling of civilian aircraft, and one of the reasons the Global Hawk is cleared to fly over civilian airspace. Abruptly, the 'crossover' between two of the military satellites used to guide the Hawk failed, due to human error.
A Predator drone prepares for take-off
This was what its autonomy software had been designed for - 'contingencies' are programmed into its software so that it can respond to unforeseen events. The Hawk turned itself around, entirely without satellite guidance, and returned to the airbase it had flown from. Ten minutes later, it landed at the base.
'Our first fully autonomous landing was in 1975,' says Dane Marolt, international business development director of the RQ-4, a former pilot who has overseen Northrop Grumman's autonomous drones programme since it first began.
'The RQ-4 is totally autonomous. It is a mouse-click aircraft. But there is no pilot flying this. As it stands, the U.S. Army and Navy choose not to use it in this way - there is a pilot in command.'
Every weapons company says the same thing - that it is their computer software that gives them the edge. The equipment inside the UAVs may not be cutting-edge, but the software is. And this software isn't as easy to protect, or to copyright, as a vehicle. It's also much more easily copied. Hezbollah has already fired captured drones back at Israel from the West Bank. There are other risks, too - last year, insurgents hacked into the video feeds of Predator drones flying over Iraq.
The website DIY Drones is a thriving community of do-it-yourself drone builders and operators, building drones that look eerily similar to - or are copies of - the weapons employed currently by the West. For a terrorist, or a lone psychopath, the idea of a vehicle that could launch, find targets and attack autonomously must seem like the ultimate risk-free weapon - a suicide bomb without a suicide bomber.
Tribesmen gather at the site of a missile attack by a U.S. drone in Pakistan, which killed up to six people in 2008
Autonomy is far more ubiquitous than people think, but it brings with it problems and dangers. The AEGIS shipboard computer used on board American destroyers controls their anti-missile systems. It works so quickly that operators simply tell it whether to shoot fighters or bombers first when the ship comes under attack - the ship then acquires targets and shoots on its own. It was an AEGIS system that had been left in attack mode that shot down an Iranian airliner in 1988. The system had mistakenly identified it as an enemy fighter.
But while mistakes of that magnitude are rare, a report, The Year Of The Drone, by the New America Foundation, an American non-profit think tank, has analysed drone strikes against militants in Pakistan and has found that the level of civilian casualties was such that it undermined any claims of drones being 'precision' weapons. The use of weaponised drones might have reduced the number of captured pilots - but their capacity to strike precisely is questionable.
'Our research shows that some two-thirds of those killed in the strikes since 2004 have been described as militants, implying a civilian casualty rate of about one-third,' says the foundation's Katherine Tiedemann.
Philip Alston, the UN's Special Rapporteur for Extrajudicial Executions, alleges that the use of drones over states such as Pakistan, with which the US has not declared war, 'might violate international humanitarian law and international human rights law'.
President Obama's State Department legal adviser Harold Koh replied to Alston's allegations saying, 'Our procedures and practices for identifying lawful targets are extremely robust, and advanced technologies have helped to make our targeting even more precise.'
Earlier this month, a drone strike on Boya village, in Pakistan's North Waziristan, killed between three and five Al-Qaeda militants, according to reports, but also up to 13 civilians. Human Rights Watch is trying to open debate on the use of the weapons in Afghanistan and Pakistan. The statistics in The Year Of The Drone allege that more than 400 civilians have been killed by drones in Pakistan in just one year - and its authors allege that the U.S. government is not open about the casualties.
'The closest a government official has come to publicly recognising the civilian casualties is an anonymous quote suggesting that only 20 civilians have been killed in drone strikes in Pakistan in the last two years,' says Tiedemann. But still the drive among the West's armies is to allow not more control over UAVs, but less.
Boeing, like BAE, is already developing unmanned aircraft that will go well beyond its current roles of surveillance and attacking ground troops. Boeing and BAE's unmanned planes look and operate like stealth bombers - a role in which communication with the outside world is likely to give away a plane's decision.
At BAE the black, triangular shape of Tanaris looks instantly familiar - it's almost identical to the American B-2 stealth bomber. It's a UAV designed for a different kind of warfare - not against tribesmen armed with AK-47s but against modern nations equipped with radar, satellites and electronic counter-measures. To maintain full stealth cover, it is capable of severing communications with its handlers and travelling without radio contact for up to 36 hours.
Tanaris is a so-called 'black project' - it's introduced as a model in a room at BAE's headquarters in Preston, and the three senior managers who introduce it are deliberately vague about where Tanaris might be used, what weapons it might carry, or any context in which it might be deployed. Tanaris will take its first flights next year and is a 'test-bed' for future technologies. Some of the technologies inside Tanaris will be used in MoD vehicles until 2025.
'One of the critical ways UAVs will improve is by staying up in the air longer - current models can only remain airborne for around 80 hours,' says the University of Reading's Kevin Warwick.
'The American military research organisation Darpa has put out a contract called Vulture looking for a solar-powered UAV that can remain airborne for five years. On the more micro scale, UAVs will have a role flying in and out of buildings. They'll also continue to become more autonomous. "Drone" makes it sound quite friendly and politically digestible. These aren't drones. They're hunter-killers. Other systems in development might work as "swarms", communicating with one another to carry out the mission.
'That's the worry - they make the decisions. What are these decisions? If it's against the enemy, it's fine - but what happens if it decides that I'm the enemy?'
By Rob Waugh
0 comments:
Post a Comment