An army of sewer robots could keep our pipes clean, but they’ll need to learn to communicate

Pipebots will swim around the network of sewage and clean water
pipes.Human Studio, Author provided

Hidden from sight, under the UK’s roads, buildings and parks,
lies about one
million kilometres
 of pipes. Maintaining
and repairing
 these pipes require about 1.5 million road
excavations a year, which causes either full or partial road
closures. These works are noisy, dirty and cause a lot of
inconvenience. They also cost around £5.5 billion a year.

It doesn’t have to be this way. Research teams like mine are
working on a way of reducing the time and money that goes into
maintaining pipes, by developing infrastructure robots.

In the future, these robots will work around and for us to
repair our roads, inspect our water and sewer pipes, maintain our lamp
our bridges
 and look after other important infrastructure. They
will be able to go to places difficult or dangerous for humans,
such as sewer pipes full of noxious

We are developing small robots to work in underground pipe
networks, in both clean water and sewers. They will inspect them
for leakages and blockages, map where the pipes are and monitor
their condition for any signs of trouble. But what happens when the
robots need to go to places where our existing wireless
communications cannot reach them? If we cannot communicate with
them, we cannot stay in control.

The pipe bots

The underground pipe networks are complex, varied, and difficult
to work in. There are many different pipe sizes, made of different
materials, placed at many different depths. They are connected in
lots of different configurations and filled to different extents
with different contents.

Pipebots is a large UK
government-funded project working on robots that will help maintain
the pipe system. These robots will come in different sizes,
depending on the pipes they are in. For example, the smallest ones
will have to fit in a cube with a side of 2.5cm (1 inch), while the
largest ones will be as long as 50cm.

They will operate autonomously, thanks to the array of sensors
on board. The robots will use computer vision and
a combination of an accelerometer, a gyroscope and a magnetic field
sensor to detect where they are. They will have ultrasound
and infrared
distance sensors
 to help them navigate the pipes. Finally, they
will also have acoustic and ultrasound
 to detect cracks in water pipes, blockages in sewer
pipes, and to measure the overall condition of these pipes.

The information gathered this way will be sent to the water
companies responsible for the pipes. In the first instance, the
robots will just monitor the pipes and call in a separate repair
team when necessary.

One of the biggest challenges will be making them communicate
with each other through the pipes. This requires a wireless
communications network that can function in a variety of conditions
since the pipes might be empty, full of water or sewage, or
somewhere in between. The three main options we are exploring are
radio waves, sound waves and light.


The robots will have acoustic and ultrasound sensors to detect
cracks in water pipes. Human Studio, Author provided

Radio waves

Wireless communication technology using radio waves is
everywhere these days – wifi, Bluetooth, and of course, mobile
phones networks such as 4G. Each of these work at a different
frequency and have different bandwidths.

Unfortunately, none of these signals can go through soil and
earth, we are all too familiar with how mobile phone connection
drops when a train goes through a tunnel. However, if we had a base
station already within the tunnel, it would allow radio waves to
travel along its length.

Thankfully, sewer pipes look a lot like tunnels to radio waves
– at least when they’re relatively empty. We are
adapting technology similar to
wifi and Bluetooth to make sure the sewer inspecting Pipebots
always have a connection to the control centre.

Water blocks radio waves even more than soil and earth. In fact,
at high enough frequencies, it acts as a mirror.
So to stay in control of the robots in our water pipes, we have to
use both sound and light.


They will have ultrasound and infrared distance sensors to help
them navigate the pipes. Human Studio, Author provided

Sound and light

Wireless communication methods using sound and light are not
widely commercially available yet. But they are making waves in the
research community.

One method, visible light
communication (VLC),
 uses transmitters and receivers, such as
LEDs, that are small and energy efficient, and also provide
dazzling data rates, on the order of tens of gigabits per
. However, because light travels in a straight line, VLC
can only be used when robots close to each other need to
communicate. One potential solution is to have many robots in the
same pipe, forming a daisy chain along which information can travel
around bends in pipes.

on the other hand
, can travel for miles along pipes and will
travel around corners with ease. The downside is speakers and
microphones can be power hungry, and sound does not offer
particularly high data rates. Instead of the several billion bits
per second
 that can be sent using 5G and post-5G technology,
sound waves can only carry a few bits per
. While this will be enough to know if a particular robot
is still functioning, it will not be enough to relay a lot of
useful information about the pipes.

The first pipe robot in action

It won’t be a case of picking either radio, sound or light
waves. The wireless communication network we are developing for our
subterranean little helpers will use a combination of these
methods. This will ensure the robots do what they are supposed to
do, that we stay in control of them, and they deliver on their

We hope to have a full Pipebots system demonstrated in a
realistic network by 2024. Once that is successful, we’ll need to
go through a thorough certification and compliance process to
ensure that Pipebots will be safe to deploy in live water and sewer

Originally written by
Research Fellow in the School of Electronic and
Electrical Engineering, University of Leeds | January 26,
for The

original article