IEC E-Tech
article here
As the market for drones rapidly expands, standardization
organizations and technical committees need to cooperate to avoid duplication.
Unmanned Aerial Vehicles (UAVs) – drones in other words –
are becoming ubiquitous across multiple industrial and commercial sectors. Some
pundits predict that by 2030 they will be as common in the sky as cars on the
road. While estimates of market size vary, researchers agree on the accelerated
growth of drones over the next decade. Forecasts of global UAV market capital
range from USD 57,8 billion by 2030 to USD 68,64 billion by
2034.
Their utility is expanding in applications as diverse as
capturing previously unattainable shots for filmmakers, aerial analysis and
fertilization of crops, mail delivery or even airborne taxis, but there are
several technological developments which are driving adoption.
Chief among these is the use of artificial intelligence
(AI), integrated in a variety of ways to boost performance and capabilities.
There are also developments in the ways to power drones to improve their main
limitation, which is flight time.
AI improves precision
While drone technology has existed for decades, it has
mainly, for reasons of cost, been concentrated in the hands of the military.
That changed when lower-cost lightweight drones carrying cameras emerged around
2010. The launch of the Phantom series of drones by a Chinese
developer from 2013 onwards catalyzed the consumer and commercial UAV market.
While high-quality image capture from these drones took off
– especially in industries like filmmaking – “the maturity for other
applications such as mapping, surveying and inspection was very low,” explains
Hendrik Bödecker, CFO of a Hamburg-based drone industry market researcher
and consultancy firm. “There was very limited software available to analyze
this data. For applications such as mapping or surveying, you needed
optimization software to extract accurate information.”
This has changed dramatically in recent years with AI and
machine learning (ML) being used to analyze large volumes of data gathered from
sensors, for example in mapping, spectral analysis and infrastructure
inspections. “If you collect thousands of images during a bridge or power line
inspection, traditional methods would require inspectors to manually review
everything. Today, ML models can identify anomalies such as corrosion or
structural defects based on trained image datasets,” Bodecker explains.
In precision agriculture, for example, the UAV, sensor and
software package of one US developer can analyze up to 15 spectral
bands over farmland. “AI can dramatically speed up post-processing, identifying
where crops need more water, fertilizer or pest intervention,” explains
Bill Irby, the company’s CEO. “Removing humans from the analysis loop allows
for faster decision-making, which improves yields and supports global food
security.”
Applying AI to video streams for surveillance or analysis
“can significantly enhance advanced target recognition, helping operators
identify and classify objects on the ground more quickly and efficiently,” Irby
adds. Ten years ago, mapping accuracy might have been around ten centimetres.
Today, thanks largely to software improvements, accuracy can be one centimetre
or better.
A claimed world’s first AI-powered drone combining
optical and electronic fog-penetration technologies is claimed to boost clarity
by up to 50% in rain or fog. However, Bödecker notes that this type of AI “is
really software-based data analytics and is independent of whether the data
comes from a drone, a crane or another platform”.
The IEC and ISO have formed a joint technical
committee, SC 42, to standardize multiple AI requirements and
applications. One of its standards series, ISO/IEC 5259, deals with the
quality of data for analytics and machine learning.
Autonomous flying not ready for take-off
Software development is also a key enabler for programming
flights and controlling drones remotely, as long as there is connectivity. For
beyond-visual-line-of-sight (BVLOS) operations, users must fully trust the
programmed flight path from point A to point B. This is where the first major
AI challenge appears and it has to do with unmanned aerial systems (UAS) for
controlling the drone in the air. Some companies claim to have intelligent
onboard communication or decision-making, but this is often limited to basic
computer vision, such as obstacle detection. Even then, a drone may not be able
to decide independently whether to move left or right in complex environments.
“Many companies claim to have ‘AI drones,’ but that is not
really true,” says Bödecker. “Drones are not yet capable of autonomously flying
from A to B beyond visual line of sight without human oversight. If you are
inspecting a tower beyond a mountain, for example, you still need connectivity,
visual contact or a human operator. The idea of fully pilotless commercial
missions is not yet a reality – it remains hype.”
The industry is moving toward autonomous, pilot-free
operation, but according to Irby, there will always need to be a human
overseeing operations. Even so, AI can help. “Whether in Europe, the US or
elsewhere, AI tools can help drones navigate approved flight corridors and
remain within permitted parameters such as altitude and speed, while avoiding
restricted areas,” Irby says. “That kind of capability would greatly enhance
global drone operations.”
One key concept is “one-to-many” control, where a single
operator manages multiple drones. AI enables that by handling navigation,
compliance and routine decision-making. “From a business perspective, this
increases efficiency and reduces costs,” Irby continues. “That said, regulators
like the Federal Aviation Administration in the US (FAA) or the European
Aviation Safety agency (EASA) will likely continue to require a human pilot
ultimately responsible for safety. It will be a long time before pilots are fully
removed from the loop.”
Technology that is autonomous can be confused with automated
tech. So-called swarm or teaming technology is one example. In theory, swarm
intelligence would involve drones dynamically communicating and coordinating
with each other in real time. This is not the same as pre-programmed drones (as
used in light shows) where hundreds or thousands of drones are
coordinated.
“There are proof-of-concept demonstrations, but no widely
deployed, proven systems,” says Bödecker. “Regulation is also a major barrier.
For now, swarm technology remains largely hyped outside of limited military
research.”
How to solve the flight time challenge
Flight time is widely considered one of the main limitations
of drone development. This is closely tied to battery technology, weight and
power efficiency. Most drones – over 99% – are battery powered (typically
lithium-ion), estimates Bödecker.
The issue with batteries is power density. The more energy
you can pack into the same weight, the longer a drone can fly. The challenge is
always balancing the added weight of batteries against performance gains.
Irby claims his systems currently fly for up to 90 minutes.
If battery density were doubled, flight time could theoretically double as
well. Most drones (around 97% according to Bödecker) use multi-rotors and
operate similarly to helicopters. They can take off and land vertically, then
move horizontally once airborne making them a good option when flying in
tight spaces. However, “multi-rotors consume a lot of power, especially for
hovering and take-off, which limits flight time,” Bödecker says.
With a fixed-wing aircraft, once in forward motion, the
wings generate lift through airflow and pressure differentials. According to
Irby, whose products use a fixed-wing design, this “dramatically improves
efficiency”.
Nonetheless, Bödecker is sceptical of most manufacturer
claims. In real-world conditions (when factors like wind drag and payload are
taken into account), most drones achieve around 25 to 35 minutes of flight
time, he says, even if manufacturers claim up to 50 minutes under ideal
conditions. In practice, this limitation is often acceptable because many
commercial use cases do not require longer flight times.
For applications such as power line inspections, longer
flight times could be useful. In these cases, drones powered by two-stroke
combustion engines can fly for up to two hours and carry heavy payloads.
However, they are expensive – often costing over EUR 100 000 – require
significant maintenance, are very loud and face regulatory challenges due to
weight and noise.
Standards for hybrid systems
Hybrid systems combining batteries and/or solar photovoltaic
(PV) cells with combustion engines are a solution to this: battery power being
used to reduce noise during take-off and landing before switching to a
combustion engine in flight.
Recognizing the need for specific guidance documents in this
area, the technical committee which prepares standards for solar PV systems has
formed a project team, IEC TC 82 PT 600, on
vehicle-integrated photovoltaic systems (VIPV) and is planning to develop two
new technical reports in this area.
The safety and performance of lithium-ion cells used in
batteries are standardized by the IEC. The IEC 62660 series on
secondary lithium-ion cells for the propulsion of electric vehicles (EVs) is a
three-part series which covers performance testing, reliability tests and
safety requirements. EV battery packs themselves are standardized by ISO
124052.
Hybrid fuel- and electric-powered designs can be more
suitable for larger aircraft, especially those operating on conventional
aviation fuel since electrical systems can extend flight time. Hybrid power
configurations can also benefit applications requiring heavier payloads, such
as the Hydra 400, which uses electric rotors and jet turbines for lift and
propulsion of payloads up to 400 kg.
Hydrogen is a way forward
Due to the high energy density of hydrogen,
hydrogen-powered drones are another emerging technology permitting longer
flights with heavier payloads than battery technology. A claimed “world’s
first” commercial hydrogen fuel-cell power pack for drones was launched
last year in a design which retains batteries for redundancy. Meanwhile, a
Ukrainian developer behind a combustion-engine drone capable of flying for
28 hours in military operations recently reworked the design with hydrogen
fuelling an electric motor. While this cut flight times in half, its advantages
include a “negligible thermal signature”, meaning it avoids being detected by
infrared sensors.
The main challenges with hydrogen are infrastructure (for
refuelling, for instance), storage weight (think massive gas tanks) and cost.
The technology is improving, and more manufacturers are adopting it, but it
remains a niche solution for now. Some standards for hydrogen applications
which could be useful for the overall safety of hydrogen powered drones exist.
As hydrogen is odourless and invisible, leaks can be
difficult to detect and thus potentially dangerous. IEC TC 31 prepares
standards for equipment used in explosive and hazardous atmospheres. To ensure
global compliance and safety with TC 31 standards, IECEx, the IEC
Conformity Assessment (CA) System which oversees hydrogen-related
certifications, is expanding its scope relating to testing and certification in
the area of hydrogen technologies. The CA system has partnered with many other
international organizations, including ISO, establishing formal liaisons
with ISO TC/197, relating to testing and certification in the area of
hydrogen technologies, and more recently with IEC TC 105 for fuel
cells. This TC has published IEC 62282-4-202 which covers the
performance test methods of fuel cell power systems for unmanned aircraft. This
encompasses start-up, shutdown, power output, continuous running time, warning
and monitoring, and environmental compatibility among other areas.
More standards are needed
According to a report by the British Standards Institution
(BSI), there are over 650 standards applicable to UAVs/UAS, many of
which were originally developed for manned aviation and are either directly
transferable or will require adaptation. Further harmonization is needed to
fully address drone design, testing and operation.
The IEC has two technical committees, IEC TC 97 and IEC
TC 107, specifically dedicated to the aviation sector, which still have to
embark on specific standards for drones. But it is quite probable that many of
theirs standards can be adapted to this rapidly growing market.
However, “the drone value chain is perhaps the most diverse
and complicated of any industry or consortium that exists in the world,” warns
Kishor Narang, principal design strategist and architect at an India-based
consultancy which leads multiple global standardization initiatives at the
IEC, ISO, ITU and IETF, notably in the area of smart cities
and EVs. “Agreeing on common harmonized standards is a major hurdle for all of
us”, he adds.
The principal categories where standards could help shape
the sector are related to safety, public acceptance and reinforcing
regulations. Specific subtopics that need addressing include operational risk
assessment (ORA), maintenance, data capture and processing.
“Some drone stakeholders confuse regulation and standards,”
says Narang. “Globally, there is a lack of awareness of existing standards and
standards being developed across the aviation and industry sectors. A perceived
lack of clarity around safety regulations creates uncertainty around
priorities. For instance, there is a potential divergence between the US and
Europe on some key technical standards, for example for BVLOS.”
Narang also highlights duplication of standards and
best-practice activity across organizations within industry groups such as
construction. In addition, some regulations are only applicable to the outdoor
use of UAS. He stresses that “a sophisticated management system is desperately
needed as drones become more widely used. The most basic functions of a drone
management system are creating, editing and dispatching missions, keeping track
of the real-time status of each drone and storing, analyzing data in a
structured format.”
In addition to his many roles in the IEC, Narang is an
expert within a recently formed IEC System Committee which deals with
electrified transport, and some experts think it should take on the job of
coordinating standardization in that space. Whatever the outcome, the need for
cooperation between standards organizations is paramount.
No comments:
Post a Comment