Unmanned aerial vehicles: ICT on the fly
The Global Hawk provides high-resolution synthetic aperture radar and electro-optical/infrared imagery at long range
An ultimate goal is to have an entirely autonomous system
Next-generation flying eye: Draganflyer X4 with its Micro Video camera. Video is stored using the camera’s memory card
A real-time video feed from the micro-video camera can also be sent to a handheld controller below
UAVs could be used for detecting and tracking forest fires when it is unsafe to fly manned craft over hazardous remote territory
The use of UAVs in the military context is a great headline grabber, but a wealth of computer technology is really helping the sector take off in much more creative directions. How will we be employing drones in the upcoming decades?
They're being used for everything from terrorising terrorists to monitoring tuna stocks in the ocean. Children are piloting them with smartphones, and law-enforcement officials are using them to track down and capture most-wanted targets. Unmanned aerial vehicles (UAVs), more popularly known as drones, have developed rapidly in recent years, thanks to some largely untrumpeted developments in information and communications technology.
The terms 'UAV' and 'drone' cover a multitude of vehicle types, and the situation is quickly approaching where sub-definitions will be needed to keep up with the variety of products on the market.
The UAV market continues to grow rapidly; in some respects it's a bit of a free-for-all, as inventive and innovative engineers and designers launch their own peculiar take on 'drone chic'. Then there's the question of regulation, and where and when the current crop of products are legally entitled to fly - and who is liable should they fall out of the sky with disastrous results.
Americans are leading here on multiple fronts. The US Department of Defense (DoD) classifies UAVs into five groups, based on size, weight and capability. Group 1 consists of vehicles less than 20lb, flying less than 1,200ft. Group 2, between 21 and 55lb, flies below 3,500ft. While group 3 spans a far larger window, with weights up to 1,320lb. Groups 4 and 5 are reserved for the heavy-hitters: long-range, heavy-endurance craft. Whereas lower-class vehicles give you agility and the ability to hand-launch, higher-class UAVs can fly higher, for longer; but even in the lower classes, endurances are still respectable. The class two Boeing ScanEagle craft used for surveillance by the Canadian military in Afghanistan, for instance, can conduct missions of 13 hours.
As director of strategic UAV campaigns at Honeywell Defense and Space Prabha Gopinath points out, higher classes mean heavier payloads in respect to on-board technology. "These now include high-definition cameras with heavy lenses, chemical sniffers, and imaging cameras that can look through a few inches of dirt," he explains, along with sensors designed to look in various directions. Such sophisticated and powerful utilities usually come with a weight penalty that can comprise other performance attributes.
Heavier payloads can also mean weapons, of course. The Global Atomics Predator is perhaps the best-known weaponised UAV, having carried out many missions in Pakistan. As designers try to squeeze more payload into vehicles with better capabilities, they must labour over size, weight and power (SWAP) ratios. "There are efforts to shrink the size of UAVs, and the ability to do more tactical and agile flight," says Tim Chung, assistant professor of systems engineering at the Naval Postgraduate School. "There is a huge investment in understanding how to increase the power density." Researchers are working on capturing wind energy from dynamic soaring techniques and to take more advantage of thermals, Chung says: "And of course, there are solar cells along the wings." Designers are also exploring emerging battery technologies, including hybrids.
Jeremy Laliberte, assistant professor at Carleton University and project manager for the Capstone UAV Aircraft Design Project, says that the real cutting-edge work in the drone sector is happening in the lighter, smaller UAV groups. "You're going to see that divergence, where the big stuff will be like any other airplane," he predicts. "Where things get interesting is with the smaller craft. With small 20-30lb UAVs, you can start exploring more materials, like 3D printing and integrated antennas."
Other technologies such as all-electric propulsion and fuel cells become important here too. "That's the nice thing about the small UAVs: they could become a test bed for technologies that could find their way onto manned aircraft," Laliberte continues.
This technology starts with military applications before making its way into the commercial space. For years, the main uses of UAVs were military ones. Since the Second World War, UAVs have gradually developed from glorified flying bombs into more sophisticated vehicles used for surveillance and targeted attacks.
UAV pros and cons
Israel led the field in UAVs for military purposes, demonstrating their successful use in the Lebanese conflict in 1982, which in turn did much to encourage the US to invest in this field. Now the DoD's interest in UAVs is escalating. The 2012 DoD budget requested $3.9bn for UAV development and procurement – an almost six-fold increase since 2001. It will spend over $37.5bn on them over the next four years.
The appeal of UAVs for the military lies partly in saving money, and partly in saving pilots, and probably in public relations derived from both of these factors. The Northrop Grumman Global Hawk RQ4 UAV has a range of 5,400 nautical miles, and is designed to stay aloft for days, with the pilot safely ensconced at a remote ground base. However, the pros and cons of manned versus unmanned spy aircraft are not as clear-cut as they might at first appear.
Some sources claim some Global Hawks cost about $35m each, while developments can push the total price to $218m. BAMS, a proposed derivative for use by the US Navy, could cost even more.
Why do these aircraft cost so much? Part of the overhead stems from the fact that some 54 per cent of the Global Hawk's cost lies in its payload. "The drone is just a vehicle: it is the data that it can collect that's important," explains Wayne Crowe, executive director of Unmanned Systems, a non-profit consortium of UAV companies based in Canada. "You must put the sensors and data together."
It goes beyond even that. Outside the drone and the payload, the UAV relies on a communication system to connect it to ground crew that must be secure and reliable. And once it gets its data back to the ground, it must be analysed. What started out as a spy-in-the-sky is now verging on something akin to a flying enterprise ICT infrastructure, except that we are still at the stage when much of the hardware and software is bespoke (therefore expensive).
Communications systems are perhaps the most challenging. High-speed links are necessary for communicating with UAVs that may be on a mission halfway round the world, and designers must also allow for the fact that even the best communications links may break. A single Global Hawk requires a 500Mb/sec link, which is estimated to be five times more than the entire US military used during the 1991 Gulf War.
The issue of UAV regulation may come further to the fore if UAV-sourced evidence is used in criminal cases.
Some police forces are exploring using small quad-motor rotor units to get forensic aerial overviews of a crime scene. Seattle police have a 3lb Draganfly X6 DraganFlyer UAS (unmanned aerial system, as is its preferred designation) that can stay aloft for 20 minutes and beam video and thermal imaging data to operators on the ground.
Texas police are using $300,000 Vanguard ShadowHawk UASs capable of firing tazers and rubber bullets, although the weapons' capabilities have not as yet been deployed. US Customs and Border patrol officials are also using larger drones on the US-Mexico border.
Companies are emerging to target these markets. "We take technology that's funded by the military, and we can put it in law enforcement for a tenth of the cost," says Chris Miser, founder of Falcon UAV. The company offers a basic airframe for $13,000, with a $1,000 autopilot option and a ground control station for under $3,000.
UAVs meet Big Data
Surveillance and data collection is a key function of UAVs that spans all these applications. "It puts eyes in the sky," says Miser, adding that a drone can hover above a known individual's house, or be used for search-and-rescue operations.
Back-end data analysis for applications such as these is a critical part of the UAV ecosystem. In the military space, data processing becomes even more critical.
"The biggest problem is not putting better payloads in the sky," says Honeywell's Gopinath, "it is dealing with the massive amounts of data that comes from these birds." Predators can deliver a terabyte of data from their sensor package for every 60 minutes that they spend in the sky. Standard commercial database software from companies like IBM or Oracle can then be used to interpret the data, using standard relational or object-oriented techniques.
However, backlogs are still building. "How to fuse that information and disseminate it has become a major issue," explains Jeff Kline, director of CRUSER – the Consortium for Robots and Unmanned Systems Education and Research. Like enterprise IT, UAV operations face ever-escalating data sets. "We are collecting more information than we can process now," Kline also admits.
The trick involves distilling the data upstream, and displaying it so that humans are not overwhelmed. One approach to this is to process a lot of the data on-board the UAV itself. "We're trying to figure out what we will let the robot do," Kline adds, "and it's still an open question."
This processing autonomy can be implemented at various levels, starting with none at all. "You could just send down all of the raw data," says the Naval Postgraduate School's Chung. "Or you could find all the cars and send thumbnail shots down."
But things can become far more nuanced. A drone might look only for red cars, is a possibility that Chung posits: "We have many researchers looking at image processing algorithms... The question is, at which level you want a UAV to make an executive decision? At a 'red vehicle' decision level? Or where it decides that an individual [on the ground] is doing a very bad thing?"
- IET Event: ‘UAVs in Civilian Airspace’ - 22 March 2013
- www.nps.edu/research/cruser/index.html NovelSat
The Drone hack prevention measures
Bandwidth aside, securing UAV links is another challenge. Honeywell's Prabha Gopinpath outlines intrusion scenarios such as full command and control, where an attacker takes control of the aircraft; interception, where someone taps into video feeds to define the intent of a mission; and deception, where someone fools the on-board control system into thinking it is somewhere that it isn't. Radio jamming and stealing a ground station controller are also possibilities.
"The command and control links are always encrypted, and in addition to the 128bit or 256bit keys there is also a proprietary information coding scheme," explains Gopinpath. "So even if [hackers] cracked the encryption they would still need to figure-out the information encoding scheme." Using a new hash with every mission when encoding information makes it very difficult to break a stream, even if someone misappropriates a ground controller.
To prevent GPS spoofing, every military aircraft uses a range of sources, including inertial navigation, to compute a position in-flight. If someone spoofs the GPS signal, the aircraft should be smart enough to register that it could not have moved hundreds of miles in an instant, and will ignore the signal, Gopinath continues.
In the event of a jammed or lost signal, military craft revert to lost-link behaviour, in which they either return home, or do something else pre-defined, such as proceeding to the next waypoint and maintaining that position until comms are re-established. They can also fall back to real-time terrain mapping to help navigate when a signal is down.
All of this again requires more on-board intelligence. Despite the image of UAVs dependent of being under the direct control of remote, ground-based pilots and operators, an ultimate goal is to have an entirely autonomous system, according to Jeff Kline, director of the Consortium for Robots and Unmanned Systems Education and Research (CRUSER). "Autonomy is a level of degree," he says. "Unmanned cruise missiles have been able to do that for years." Non-military deployments do not face the same challenges, but that's arguably partly because the regulations forbid them from doing as much.
Balancing regulation against requirements
In the US, the FAA imposes strict rules on government agencies flying drones, and has banned the commercial use of UAVs altogether, allowing only 'experimental airworthiness' certificates for controlled training, research, and marketing purposes. Such certificates preclude the carrying of people or property for compensation or hire. Congress is pushing the FAA to introduce new rules for commercially-operated drones in domestic airspace by 2015.
In other countries such as Canada and the UK, meanwhile, commercial drones usually require a separate certificate of authorisation for each mission. In the UK, the Civil Aviation Authority released CAP 722 in 2010, an updated document that regulates the use of UAVs weighting 7-20kg. They must be flown below 400ft, and within line-of-sight.
Canada is working on similar regulations, but the process is slow. Like the US, it has vast tracts of territory with little-to-no population, which could arguably benefit greatly from some manner of scheduled UAV monitoring. They could be used for forest fires, for example, explains Unmanned Systems' Wayne Crowe, especially at night when it is very unsafe to fly manned craft over hazardous and/or remote territorial expanses. "Because fires are so unpredictable, these things can be used to monitor where they're going," he points out.
As far as the more controversial commercial applications go, geophysical surveys for industries including mining or oil and gas pipeline monitoring are also possible uses. UAVs could also be used to track remote power lines that are down due to storm damage, so that maintenance and repair crews can be directed more efficiently to the location of the problem.
More academic and humanitarian applications include wildlife migration monitoring. "These UAVs are unobtrusive, and they can fly overhead without spooking the wildlife," Crowe says. "The challenge is that you're getting streaming video coming back, and you can quickly identify what it is that you are looking for. That involves target recognition techniques." In other words, the cleverer the application, the more it costs.
Unlike the higher-grouped military class, UAVs used for non-military purposes are often far smaller, and they are used for a variety of reasons. UAVs were employed during Japan's March 2011 post-earthquake nuclear incident to collect a radiologic and visual data, for example.
Autonomy and control: why software is key to safety
Theoretically, UAVs have the potential to take on a range of tasks, such as sensing and avoiding other vehicles, which would provisionally open up opportunities for flying in congested airspace. But how reliable is the software, and how advanced are the control algorithms?
"You have two concerns with making planes reliable: one is stopping them crashing into each other, and the other is [preventing them from] crashing into the ground," says Professor Robert Dewar, CEO of AdaCore, which writes ADA-based software for the aerospace sector. ADA is an object-oriented programming language known for its ability to produce software that can be audited for reliability, and is often found in safety-critical applications.
"It's hard to imagine that this software is as reliable as it needs to be when it has not gone through rigorous certification standards," Dewar says of some UAV control systems. In the commercial space, many drones are ad hoc vehicles, developed by researchers or field engineering teams. "The guys making drones aren't under compulsion to do that. It's expensive."
The autonomy and capability of these drones is a crucial issue, both inside and outside the battlefield. The Bureau of Investigative Journalism has reported that (at least) 482 civilians may have been killed by Predators in armed UAV strikes in Pakistan alone while armed UAVs have been flying missions since 2004. "Is it intelligence mistakes? Operator mistakes?" asks Dewar. "Then there are cases of drones malfunctioning."
Drones have come down in battle zones, and there have also been reports of commercial drones being downed and lost – in December 2011 a US RQ-170 UAV that was captured by authorities in Iran was reported to have malfunctioned. However, generally, UAVs do not seem to be associated with particular reliability issues.
UAVs rely on a variety of comms mechanisms, depending on how close they are to controllers. Line-of-sight controls can be used to control UAVs in close proximity to ground teams, especially during launch and landing operations, but drones will switch to satellite comms when moving over the horizon on longer-range missions. How reliable are satellites in UAV communications? David Furstenberg, CEO of satellite data transmission specialist NovelSat, argues that satellites are in many cases the only way of controlling a UAV in hostile territory. "Satellite is considered to be one of the most robust technologies," he says. Satellites lie outside the atmospheric envelope, which makes them more reliable than many communications technologies. Nevertheless, UAVs must be designed to accommodate issues such as rain fade, when signals are hindered by precipitation and storm fronts.
Latency is the other major issue for satellite communications. A geostationary satellite sustains a 250ms lag between the sender and receiver of a signal. However, Furstenberg argues that weapons are tuned to handle such lags. It would be easier to launch guided missiles against static targets via higher-latency links, as opposed to dogfighting with fast-moving aerial targets. However, the fact remains that communications latency lag is one of the possible issues that will fall under examination in the case of UAV control malfunctions, especially when missions occur where drone and operator are, in effect, half a world apart.
In the early days of operation, General Atomics Predator UAVs were launched and landed while under control from line-of-sight operators on the ground, but reports suggest that armed UAVs, such as the General Atomics Predator and MQ-9 Reaper drones, can now be launched and landed by remote operators via satellite links too.
|To start a discussion topic about this article, please log in or register.|
"The Internet of Things used to be a buzz phrase in tech circles, but it's already so last century. Brace yourself for the Internet of Everything"
- Can we make further productive use of Wylfa Magnox Nuclear Station site at Angelsey? [01:08 pm 06/12/13]
- Thermal Power Output the Civilian Equivalent to a Submarine Nuclear Reactors [12:31 pm 06/12/13]
- PCB Design Courses [06:51 am 06/12/13]
- basic info [06:48 am 06/12/13]
- MBA advice needed [06:45 am 06/12/13]
Tune into our latest podcast