Recently OpenGov spoke to Mr Lup Yuen Lee, Chief Technology
Officer at UnaBiz[f1] ,
the exclusive network operator of Sigfox’s low-power
wide-area network (LPWAN) in Singapore and Taiwan. UnaBiz is the
first IoT-dedicated network operator in Asia to roll out a nationwide[PB2]
IoT network.
Recently, UnaBiz enabled full indoor coverage of the Sigfox
IoT network at all 4 terminals of Changi Airport. Smart solutions providers and
system integrators have developed Sigfox-enabled solutions such as temperature
sensors and other applications for the Smart Airport.
The Singapore-based
company partnered
Taipei City Government last year to build an IoT Innovation Lab. It is working
with Airbus to advance research in
digitalization of aircraft maintenance operations through the adoption of IoT
solutions. In collaboration with bike-sharing company, oBike, UnaBiz is rolling
out geolocation services for one million bikes on Sigfox Global LPWAN.
In a wide-ranging discussion, Mr Lee talked about different
types of IoT applications and their network and power requirements and shared
his views on the complexities of IoT development.
Deep vs Wide IoT
Mr Lee explained that there are two classes of IoT, ‘deep’
IoT and ‘wide’ IoT. Amazon Echo and Alexa are examples of deep IoT. Deep IoT
devices require high bandwidth and power supply.
With the Amazon devices, the voice command goes to the cloud
for processing and generating an output command. This has high computation power requirements and
hence, these devices don’t work well on a low-bandwidth network or low battery
power. As a result, they tend to stay fixed in offices or homes.
UnaBiz looks at wide IoT, which refers to devices that are very
light, battery-powered and operate on pervasive networks. They can work anytime,
anywhere in Singapore and do not rely on WiFi or the cellular network.
“We don’t think everyone will be able to afford an Amazon
Echo. It is a very powerful device but it is not cheap, because there’s so much
complexity inside it. In short, we’re just trying to be a very simple kind of
IoT network, where you press a button and it triggers the backend. The device does
not need to pair with Bluetooth or WiFi. It is simple and fuss-free, even for
the elderly,” Mr Lee added.
Subscription also tends to be easier, and there is no need
to worry about SIM cards because all the devices have a built-in ID. The ID
indicates that the device belongs to a certain company, and all messages can be
directed there, without any routing elsewhere. The packets are a mere 12 bytes,
so bandwidth requirement is limited and users are not expected to pay an
exorbitant price to use the network.
Cost-effective and
energy-efficient IoT solutions
When Sigfox was created in Europe, it was for primarily for outdoor applications, for low-power sensor kind of networks that need to send data intermittently. UnaBiz has been working on such applications that help collect data on outdoor environment, such as the weather and haze conditions. UnaBiz has been exploring indoor applications as well, such as tackling rodent infestations in F&B or retail shopping malls.
The trouble with
rat traps is that if a rat gets caught, it has to be gotten rid of immediately.
Otherwise the rodent will decompose, and the other rats will disperse. So how
can building owners know if there is a rat stuck in the trap and alert someone
to clean it up. Doing regular manual checks is simply a waste of manpower.
“The problem with
this kind of use case is that the rats run around in very strange places, deep
inside the building. You cannot guarantee that there’s WiFi network in air con
vents, ducts etc.,” said Mr Lee.
Therefore, the
solution needs to be able to penetrate into distant locations, without being
constrained to just public areas or by WiFi coverage. Sigfox was found to be a
good solution because of its pervasiveness. One base station can provide
coverage for the whole building.
Mr
Lee said, “We’re actually trying a few types of tracking solutions. You can
install a GPS module, however as we all know, running GPS on any device uses up
a lot of power.”
“The second idea
being explored is WiFi geolocation –
like an Android or iPhone which can use WiFi hotspots for locations –
but if you think about it, they might not work in the wild in areas such as a
reservoir because there is no network there, or in secluded areas such as big
drains or canals for flood monitoring, on the rooftops of buildings with solar
panels to monitor power storage and usage (UnaBiz is currently working with Sunseap on power metering), or at the basements of industrial buildings
for monitoring water leakage.”
The third alternative
would be to use the Sigfox network for geolocation.
Mr Lee said, “You
can either use a high-power one which will drain your battery faster or you can
choose something like Sigfox geolocation which requires no power, as long as it
transmits one message a day.”
“For Sigfox, it’s
easy, just one base station can penetrate the whole building indoors. There’s
no need to shift the base station around and you do not need to put in additional
base stations.”
“When we talk about networks, power and costs matter. If the
rat trap needs to be hooked up to the mains, then it’s not going to work. You
cannot be assured that there will be power source anywhere you go- so it has to
be battery-powered. Battery power means that it has to be a very low power kind
of network, WiFi will probably drain it because it consumes too much power,” he
added.
The reason why Sigfox is so energy efficient is that the way
it transmits in the form of a broadcast, sending out very small packets, as
mentioned earlier. Every time a message is sent, three packets are sent at
three different frequencies (this is called frequency hopping). When running on
unlicensed frequencies, some of the packets might get blocked. If one is
blocked, the others can still go through, ensuring that the message is
transmitted.
“Because it is ‘broadcast’, the communication is very simple
There’s no need to negotiate – 3G and WiFi networks need to authenticate with
the hotspot. They need to make sure the password is correct. After that they
need to keep the session alive, whereas Sigfox can shut down after each broadcast,
reducing power consumption,” Mr Lee said.
Therefore, Sigfox is ideally suited to applications that
need to be delivered at a very low cost, have less frequent communication
requirements, and require exceptional battery performance.
There are numerous smart cities applications that requires
such monitoring sensors where deployment need to be pervasive. If we think of waste
management, building management, critical infrastructure monitoring, and imagine
the need to put a sensor on all the fire hydrants, all AED (Automated external defibrillator)
devices, all the power meters, all of the trees in Singapore, the cost and
simplicity of deployment becomes crucial.
And how does Sigfox achieve wide penetration? Because it is
an ultra-narrow band technology. Transmissions on wireless networks are divided
into different channels. With Sigfox the communication channels are very
narrow. Each message is 100 Hz wide. Because these channels are so small, the possibility
of interference is very low.
Other networks like LoRa have the advantage of being able to
send bigger packets, but bigger packets also impliy higher risk of
interference.
Barriers to take-off
in IoT technologies
IoT technologies have been around for a while. And there is
a market for interesting applications. Even if it is a small market like
Singapore, technologies developed here can be exported worldwide.
Then what is holding back development and deployment?
Mr Lee has been an Adjunct Lecturer at Temasek Polytechnic School of Informatics & IT since April 2015, teaching and mentoring the next
generation of ICT professionals in networking and IoT technologies. From his teaching experience, he realised that working on IoT
technologies is quite difficult.
“Because it involves a whole range of skills. You have to
know about hardware. You have to know about these devices. You have to know
what powers these devices, what is the transmission range of these devices.”
“Next you have to go up in the cloud. These things will
transmit to some base station, the base station will be connected the cloud.
You have to figure out how the data goes into the cloud. Then how do you build
a cloud that can handle all these devices. It’s quite easy to handle one device
at a time, for prototyping. But to handle hundreds of thousands of devices is
quite challenging,” he explained. It is very difficult to find people with that
wide a skillset.”
Then there is the question to how to analyse all the data
from the devices
“Very few jobs in Singapore that have that kind of data processing
requirements. We are one of the first to actually do this kind of large scale
analytics. We need tools to be able to massage the data.”
Mr Lee also said that today we see a lot of devices created
just for the sake of it. These are examples of technology looking for a problem
to solve. Identifying problems is a crucial step.
He provided an example of a very real problem UnaBiz is
trying to address.
At a home for patients suffering from disabilities, some of
the residents go out for work. The officials want to make sure that they report
to work on time and that they also come back on time. It is about ensuring that
they are safe and are not getting lost. The home cannot afford to give the
residents expensive phones or trackers. Even if they do, the devices will run
out of battery when their clients do not return.
UnaBiz proposed using one of its motion-triggered Sigfox
devices. Residents can carry the device around and everytime they move, it
sends a message to the cloud. Then, an algorithm is used to do machine learning
and figure out where the person is. Being mindful of privacy concerns, the technology
is kept accurate to a radius of around 1 km, which is enough to know if the
person is safe, without pinpointing their exact location.
This is only one example of tracking solutions for
non-motorised assets. Other use case include tracking bicycles, people, pets.
The device must be affordable and accessible for the mainstream users to adopt
and benefit from.
Ultimately,
accelerating IoT development and deployment would require connecting the people
with expertise in devices and in cloud computing with the business people,
placing them all on one team. This would enable the creation of solutions with
real value, solving real-life problems.