Quadcopters
and other sUAVs are becoming increasingly popular among both hobbyists
and corporations looking to exploit their potential as delivery devices.
Bob Stone discusses the human factors and safety implications of sUAVs
and argues in favour of licensing and regulation.
Over
the past 18 months or so, we’ve been investigating how best to exploit
remotely controlled aerial and land-based vehicles in support of a
number of historical site surveys, prior to their reconstruction, using
virtual reality techniques.
A number of these sites present us
with quite significant challenges in terms of access. Some, such as the
old Yelverton Reservoir on the edge of Dartmoor, for example, are
surrounded by dense foliage or hidden man-made hazards. Others, such as
the 16th Century abandoned manor at Longstone, again on Dartmoor, are
cordoned off, awaiting renovation.
We’ve also been using small aerial vehicles – quadcopters and hexacopters
– to provide video footage of some of our undergraduate student
projects, including a recent successful dive of a remotely operated
submersible in a flooded quarry near Tamworth. As our experience has
increased, we have become more concerned about the human factors issues
surrounding their safe and efficient use.
Small, radio-controlled
unmanned air vehicles or sUAVs – vehicles with a mean take-off weight of
7kg or less, including the multi-rotor systems we’ve been using – have
attracted considerable attention from members of the press, who have
enthusiastically reported on a number of possible non-military
applications. For example, in 2013, Amazon announced a plan to launch a
new delivery service called Amazon Prime Air
which would deploy autonomous quadcopters to deliver orders to
customers’ addresses within 30 minutes. Other examples include the
restaurant Yo! Sushi, which has used sUAVs, controlled by waiting staff,
to deliver meals to tables.
Many commercial applications are true
flights of fancy, especially given the limited payload capabilities of
most of the commercial off-the-shelf (COTS) products, not to mention
their limited endurance in terms of power and all-weather performance.
Of
course, those issues may well disappear as the technology evolves, and,
judging by the different systems that are now widely available, this
evolution is gaining pace. Nevertheless, it has become increasingly
apparent that the main limiting factor with these devices is that of the
untrained human pilot, and there have certainly been incidents in the
recent past which demonstrate the end result of human error, system
failure, flying in inappropriate weather conditions and sheer
incompetence. In April 2014, a resident of Barrow-in-Furness was prosecuted
for having flown a radio-controlled aircraft in restricted airspace
over BAE Systems’ nuclear submarine facility. In the USA, a man was
questioned by the FBI for crashing a camera-equipped sUAV close to the
Bridgeport Harbour Electricity Generating Station.
Instances such
as these also worry authorities about the potential for using sUAVs in
support of terrorist attacks. In 2011, an al-Qaeda affiliate plan to
launch an attack on the Pentagon using an sUAV with an explosives
payload was intercepted by the FBI. In 2012, criminals flew a low-cost
quadcopter over a Brazilian prison fence to deliver cell phones to the
prisoners.
Unfortunately, there have also been numerous reports of
injuries, even fatalities, caused by loss of control of an sUAV. The
second half of 2013 was particularly bad, with fatalities in Texas,
Korea, Brazil and New York.
Of
course, there are legislative documents and guidelines covering sUAV
flights laid down by aerospace regulators, such as the Civil Aviation
Authority (CAA) and Federal Aviation Administration (FAA). The CAA has
published guidance that includes a chapter on human factors issues.
However, this particular chapter reads very much as if it has been
lifted out of a high-level defence human factors standard and, as is
often the case with those publications, is far too generic and rigid to
cope with a rapidly changing technological scene, let alone the needs of
the user of sUAV products. Certainly there are a number of relevant
issues highlighted within the chapter, including references to degraded
pilot situation awareness, pilot risk perception and strategies for
effective reversion to manual control for systems that support some
degree of in-flight autonomy. However, the style of wording and absence
of illustrated examples do little to emphasise the very real dangers of
flying under direct-view or first-person view (FPV) conditions to the
increasing population of casual or research sUAV users.
Having now
experimented with a number of sUAV solutions, we find ourselves in
strong agreement with those calling for regulations to be strengthened.
Indeed, we have recently produced our own standard operating procedures
to go some way to fill in the gaps. Over the past 12 months alone, our
hexacopter has evolved from a 2kg to 4kg payload capacity, enabling our
students to conduct research into a wide range of new payload options,
from new automatic vision and non-visual sensors, to head-slaved FPV
camera pan-and-tilt units feeding stereoscopic video data to a
head-mounted display. The addition of a gimbal-stabilised camera
platform certainly supports the generation of very impressive video
footage when undertaking FPV flying, but, on occasions, even our
experienced pilot has been “drawn in” by the stunning picture quality,
only to be alerted, just in time, by the appearance of the sUAV
propeller blades in the field of view as the vehicle becomes
progressively unstable.
It is certainly possible to undertake
flight and ground training courses for flying fixed and rotary wing
sUAVs that culminate in basic competency qualifications. For example, a
small number of organisations offer the National UAS Certificate for
Small Unmanned Aircraft and the Remote Pilot Qualification - Small
Unmanned Aircraft, both of which are recognised by the CAA as evidence
of pilot competence. However, these courses tend to be focused on those
wanting to use UAVs for professional or commercial work, as opposed to
the hobbyist or academic researcher. It is clear that, as the stories of
injuries, fatalities, property damage, invasion of privacy, trespass
and airspace incursion increase, the situation has to change, even
though any change will, without doubt, be very unpopular with many
hobbyists and retailers. Some COTS sUAV developers, for example, are
already developing firmware modifications to help prevent incursions
into restricted airspace. This is certainly a promising start, even
though it may not be met with enthusiasm from some sUAV users. However,
we believe that much more needs to be done, including ensuring that
every sUAV sold needs to be registered and marked, perhaps even chipped,
in some way so as to uniquely identify the platform with its owner.
There is no doubt that this will be difficult to enforce, especially as
3D printing technologies are increasingly being used to ‘manufacture’
replacement components.
We also believe that some form of licence
should be held by sUAV users, granted only after passing a basic
competence assessment. This could be a test similar to those used to
train and test decision-making and observational skills for driving
students. sUAV simulator packages already exist, but many fall short of
teaching and evaluating the essential skills and awareness necessary for
safe sUAV deployments. With early attention to human-centred design and
a subsequent focus on delivering appropriate sensory and behavioural
fidelity, simulators would be more than capable of testing reaction
times, manual control stability, appropriate decision-making for
flights, pilot distraction effects, and so on. Such tests would not
replace the more professional courses but would help to ensure the
growing community of COTS sUAV users use their product safely and
responsibly.
Some may complain that these measures are over the
top and totally unnecessary for something that is, after all, just a
smaller version of a radio-controlled helicopter. However, just type
“quadcopter injuries” into Google Images and make your own mind up…
By Professor Bob Stone, Director of the Human Interface Technologies Team at the University of Birmingham.
This article first appeared in Issue 530 of The Ergonomist, August 2014.
Back