Software versus hardware
However, while building a robot that looks like
a human is relatively straightforward, getting
it able to function like one is decidedly less
so. Human bodies are essentially extremely
advanced machines, capable of a wide
range of tasks of varying skill levels and able
to adapt. While one of the problems to be
surmounted in the quest to achieve either
autonomous or remotely operated robots lies
in the complexity required from the software
(and this is where AI and machine learning
come in), the difficulty of building a robot that
is physically capable of carrying out the same
types of tasks as a human being is one that
can be easily overlooked by those not familiar
with the details.
One of the major challenges in the field of
robotics today is dexterity. Manual manipulation
of objects seems simple to us, as seen every
time we open a door, or pick up a piece of
paper from a flat surface. For robots, this task
is significantly more complicated. Another
surprisingly complicated task is walking, which
is far less simple than those of us who do
it every day without thinking might realise.
Rather than repeating the same sequence of
24
FEBRUARY 2018
Walking is a surprisingly complicated task for humanoid robots.
But this is still a relatively straightforward
application, with the Hadrian performing
almost like a giant 3-D printer: programme in
the design and materials, put it in place, and
watch it go. The next generation of robots
will incorporate AI and machine learning, as
well as (in some cases) AR and VR to achieve
feats that static programmable robots are
incapable of. And these are the robots that
will have the biggest impact on the mining
and construction industries.
Of this next generation, a specific
subset is that of humanoid robots — a
field that is seeing significant research and
development, partly because of our inherent
attachment to the concept of things that
look like us, but also partly because of how
versatile such robots would be. Think about
it: a (sufficiently advanced) humanoid robot
would be capable of the same range and
type of tasks as a person, but with greater
precision and strength, and far less risk.
Certain areas of construction and mining
are inherently high risk, such as lifting,
tunnelling, or any type of deep excavation,
and humanoid robots could be employed
to improve safety and reduce the risk of
harm to human life. They could also be used
to complement or replace large, unwieldy
machines, making work in confined spaces
simpler than ever before.
TECHNOLOGY
On 16 November 2017, US robotics company Boston Dynamics uploaded a video to YouTube of its
humanoid robot Atlas doing a backflip.
movements, we are constantly making tiny
adjustments based on surface resistance,
balance, incline, and many other factors.
The difficulty of meeting both of these
challenges was on display at the biggest
robotics challenge in the world, specifically
focused on humanoid, remote-controlled
robots: the DARPA Robot Challenge, which
ran from 2013 to 2015. The competition was
inspired by the disaster at the Fukushima
Daiichi Nuclear Power Plant in Japan in 2011.
Employees were unable to open valves and
release the steam that could have averted an
explosion because they could not get close
enough in time due to massive amounts of
radiation. If robots had been on hand, this
problem could have been surmounted.
Unfortunately, not only were the eight
initially envisioned tasks — get into a standard
human vehicle and drive it to a specified
location; get out of the vehicle and travel
across rubble; clear obstacles from a doorway;
open the door, and enter the building; find a
leaking pipe and close the associated valve;
reconnect a hose or cable; climb a ladder;
and grab a tool from the site, break through a
concrete wall and exit — significantly modified
by the time of the final competition, most of
the robots struggled to complete even the
more basic course. Of particular difficulty?