Report: Building the Intelligent Campus – the Internet of Things

The final published version of this report is now available from the Jisc website

TL;DR: We’re seeing a perfect storm of innovation as pervasive connectivity and powerful miniaturised computing hardware means that almost anything can be a connected device. From toothbrushes to toasters, chairs to cupboards, venture capitalists are currently carrying out a world-wide experiment to see which of these use cases really add value. In this report I’ll look at how we could use Internet of Things devices to help us make our campuses intelligent, and consider potential perils and pitfalls


Learnometer – using IoT to improve learning []

1. Computers are talking to each other behind your back

Take a moment to look around, and imagine that all the objects you see are connected devices. What do you think they might be saying to each other? It’s really all about sensors – the hardware that gives these tiny computers ‘senses’. Some of these senses will be familiar – touch, sight, hearing. Others are ones that we would find it harder to relate to. Let’s see…

Taste: taste sensor, aka electronic tongue

Touch: touch sensor, pressure sensor

Sight: visible light camera, infrared camera, motion sensor, Light Detection And Ranging (LIDAR), light sensors

Sound: microphone, specialised decibel meter, loudspeaker, beeper, piezoelectric speaker

Smell: electronic nose, carbon monoxide, carbon dioxide, particulates (PM2.5, PM10), volatile organic compounds, dust, smoke, gas, water quality, temperature, humidity, barometer

Motion: Accelerometer, strain gauge, force gauge

Location: Global Positioning System (GPS), Global Navigation Satellite System (GLONASS), BeiDou, Galileo – also Bluetooth and Wifi triangulation and RFID proximity

Thought: Electroencephalogram (EEG) headsets have been used in medical diagnosis for decades but are now starting to become available at a consumer price point, e.g. the Emotiv Insight

Many of these sensors are quite useful in their own right, but it’s when you put them together that things start to get really interesting. Picture a robot dog (like Spot from Boston Dynamics) that sniffs out and warns you about harmful pollutants. Or a fleet of self-driving cars that scan millions of data points around them every second using LIDAR, and automatically form a flock when travelling on the open road

It’s easy to talk about connecting anything and everything to the Internet, but why are we increasingly having this conversation? The Internet of Things (IoT) has been a buzzword for many years, but without making an appreciable impact. However, two key elements have come together in the last couple of years to finally make IoT truly viable:

  • Pervasive connectivity
  • Commodised computing

In the next two sections of this report, I’ll look at each of these in turn, then I’ll go on to consider some key issues around the Internet of Things. I’ll close by introducing Jisc’s Intelligent Campus project, which is looking at research and education applications of IoT


2. Pervasive connectivity – putting every Thing on the Internet

Raspberry Pi based LoRaWAN base station

When Google revealed their prototype Chromebook in December 2010, pervasive always-on connectivity was still relatively unusual, and the idea of a computer that only worked while connected to the Internet seemed quixotic at best. It’s telling that by 2018 we now expect and assume that we will have pervasive connectivity everywhere, and many of us start to experience something approaching withdrawal symptoms when we are unable to connect to our usual systems and services

Of course the truth is that there are still plenty of places where connectivity is sporadic at best – even in developed countries like the UK. And what do we really mean by connectivity? Wifi and Bluetooth only work over quite short distances and require dedicated infrastructure installing. Furthermore, their radio signals are easily blocked by modern building materials like foil backed plasterboard. Mobile phone technology has widespread penetration, but requires contracts and subscriber information module (SIM) cards for serious use, and rural areas are all too often still not spots, even in the UK

What’s a Thing to do? It turns out that a surprisingly large number of IoT devices simply use Bluetooth or Wifi and hope for the best. This can lead to all kinds of fudges and bodges, such as using a phone or a ‘mifi’ device to provide a wireless connection to an IoT device. This is all set to change over the next few years, with the emergence of standardised approaches to connectivity for IoT devices wherever they are. The only catch is that, in a scenario that will be familiar to anyone who remembers VHS and Betamax, there are three would be standards:

Narrowband IoT (NB-IoT) – this approach uses a subset of the existing 4G LTE mobile phone comms standard, and is backed by major telecommunications companies. With NB-IoT your IoT device most likely communicates with the same masts and base stations as your phone

LoRaWAN – a ‘DIY’ approach to operating IoT networking. Anyone can set up a gateway, which works like a wifi router using unlicensed radio spectrum to communicate with IoT devices. It can service connections devices for distances of up 15km at very low power rates

Sigfox – Sigfox is a bit like a corporate version of LoRaWAN. Whereas anyone can set up the back end infrastructure for LoRaWAN based devices, and community initiatives like The Things Network exist, Sigfox licenses national operators to provide a fully supported commercial service

It’s still uncertain which if any of these three will emerge as the winner, or whether we will end up with a mixed economy. For this reason, you will see that many IoT manufacturers are hedging their bets, producing devices that have swappable radio modules, e.g. using XBee compatible sockets. However, IoT software often doesn’t have an abstraction layer for the connectivity technology, and switching from say Sigfox to LoRaWAN would require software to be rewritten and devices in the field to be reflashed or replaced

3. Small is beautiful – commoditised computing

Arduino based pollution sensor

Arduino/LoRaWAN based pollution sensor

When we consider small low power computers, many people’s thoughts would immediately turn to the Raspberry Pi and other single board computers like the Intel NUC. With price points from £5 to £35, the British built Raspberry Pi is an inexpensive and surprisingly powerful general purpose computer, with dedicated General Purpose Input/Output (GPIO) connectors that make it easy to connect to sensors and other additional hardware. The Raspberry Pi is based on the Acorn Risc Machine (ARM) chip found in most mobile phones, and is essentially the long term evolution of the BBC Micro of the 1980s

Whilst systems like these often appear in IoT environments, they are actually quite heavyweight for applications which simply require a sensor reading to be taken and reported. Consider for example noise and light levels in a room, particulate readings alongside a busy road, or infrared motion sensor. For these sorts of use cases not only is the computing power of a Raspberry Pi unnecessary, but the corresponding power consumption is unhelpful. The ideal IoT sensor device would run on battery for months if not years at a time. This requires a simpler processor than the ARM, known as a microcontroller, which only runs one program at a time

It’s very common to find microcontroller based IoT devices that are compatible with the open source Arduino technology. Most Arduino devices are programmed in a variant of C or C++, but some provide alternative coding environments, such as the Python based LoPy. The popularity of the Arduino hardware and software stack makes it easy to ‘pick up IoT’ as a developer, with huge amounts of online information available. The more advanced connectivity topics discussed above can be difficult to find reliable information on, although you will find (for example) tutorials on using LoRaWAN and even authenticating to eduroam wifi

4. IoT bear traps and tar pits

Let’s picture that you are an IoT developer. Perhaps you’ve built a Raspberry Pi powered device that connects via wifi, or an Arduino based device that uses LoRaWAN to report back sensor readings. Great – well done! There is one small catch, though – where do you send your data, and what format do you send it in? Do you need to encrypt your data, or take any other security precautions?

You might not feel that these things are important for a prototype or trial, but they quickly become issues when you want to get products from different vendors to communicate with each other, or just to produce something that stands a chance of being future proof

Here are some of the key issues you’ll face when deploying IoT:

Long range or high throughput? Wifi and Bluetooth are high speed, but only work over a short distance. IoT technologies like Sigfox, LoRaWAN and NB-IoT work over long distances but are slow and only permit small amounts of data to be exchanged. You could put 3G or 4G mobile phone SIMs into your devices, but this can be expensive

We recently heard about a Polish environmental charity that received a large phone bill after losing the SIM card from the GPS tracker it had tagged a stork with. It transpired that someone had found the GPS tracker, removed the SIM, and started using it in their own phone. Conversely, dedicated IoT networking technologies have very low data transfer speeds, which could pose some problems keeping your devices’ software up to date

Black box or open source? Some IoT solutions are presented as ‘black boxes’, with all of the technical details hidden away. Others are based on common open source software like Arduino and open standard protocols like the ISO standard Message Queueing Telemetry Transport (MQTT or Mosquito). The more open software and data formats are used, the fewer mysteries there are about how the product works

This is important because the more you know about the workings of the system, the more confident you can be about whether basic security precautions have been taken. It’s not uncommon for IoT devices to send data unencrypted (making it vulnerable to snooping) and to have back doors for remote access. US election voting machines turned out to have a default remote access password of abcde, and a similar problem with default passwords led to attackers being able to commandeer a botnet army of 1.5 million CCTV cameras

Handling personal or sensitive data. The data you’re gathering might have a clearly personal element, such as photographs of students used for sentiment analysis via Microsoft’s Face API, or even gait analysis. But there’s other data that can identify people at one step removed, such as Automated Numberplate Recognition (ANPR) scans, and the wireless Media Access Control (MAC) address of their devices

Transport for London recently trialled a system for tracking London Underground users by their MAC address, which it used to conduct research on travelling patterns – the MAC address information was protected using a cryptographic hash function

Do you actually need to own / control the IoT devices? You don’t need to create and deploy your own network of devices if you can piggyback off existing devices. Smartphones nowadays tend to have an exotic array of sensors including light and sound but also more esoteric ones such as barometers and compasses – and even air quality monitoring. You might find that you are building an app rather than a device!

5. Chips with everything? IoT and the Intelligent Campus

Jisc's Intelligent Campus architecture

Jisc’s Intelligent Campus architecture

Picture a room that knows not just how many people are in it, but also where they’re sitting, and how they’re feeling – or even what they’re thinking. Now imagine that this data is part of a ‘campus brain’, an AI that is constantly matching and classifying its data inputs and either taking action autonomously or alerting its human minders when it thinks action needs to be taken

This might sound far fetched, but in many ways we do these sorts of things already – often quite slowly and labour intensively. For instance, student feedback forms might indicate that there is a problem with a particular lecture theatre – perhaps the lighting isn’t good enough, or there is a lot of noise from a lab next door. At Jisc we have been working with a UK education technology (edtech) startup called Learnometer. The Learnometer team has built a low cost IoT device that monitors a wide range of environmental data, collating this in a cloud based analytics engine and using visual alerts when key thresholds are exceeded, such as carbon dioxide (CO2) levels

Building management systems are often set up to adjust ventilation and other climate controls in real time according to thresholds, and incorporate some of the sensor technology we’ve looked at, but they don’t tend to join up with other data sources. For instance, car park occupancy sensors might be able to tell you where there are free spaces, but the car park itself doesn’t know anything about where people are going after they park up

Let’s imagine that we connect up all these disparate data sources – suddenly we have the opportunity to make a connection between wellbeing, learning outcomes and staff performance  and the environment that our people live and work in. Perhaps there’s a correlation between a particular lecture theatre and poor learning outcomes across a range of cohorts? Perhaps there are a whole load of people who usually struggle to find a parking space, or who get stuck in traffic en route to campus?

And herein lies the rub – we’re increasingly talking about very personal information, such as learning outcomes, attendance and location data. The recent scandal around how Facebook data was abused for political purposes shows us just how significant the impact can be on people and society as a whole when sensitive data is either not being managed carefully, or when it has been released to individuals whose intentions are less than wholesome

Jisc Intelligent Campus guide

Jisc Intelligent Campus guide

At Jisc our Intelligent Campus initiative is looking into the potential of integrating IoT sensor data with other systems and services around the institution, from virtual learning environments to timetabling systems, building management to attendance monitoring. We think significant opportunities arise when a joined up approach is taken to data. Read more in our Guide to the Intelligent Campus, and do get in touch with me or James Clay if you would like to get involved in this initiative






This is a draft report – we’d love to know what you think of it, so please do leave a comment below, or get in touch. Be sure to read our other horizon scanning reports and keep a look out for new material – we aim to put a new report out every couple of months

About Jisc

We are a not-for-profit company owned by and serving the HE, FE and skills sectors in the UK. There are around 18 million users of our shared services like the Janet network, eduroam, JiscMail, and our shared data centres. We also do a number of national deals with IT vendors and publishers, such as Amazon, Google, Microsoft, Elsevier and Springer. Our other key activity is to provide advice & guidance to the sector on digital technologies in research and education

My role at Jisc is to lead a small group which is exploring the potential in research and education of emerging technologies like Virtual and Augmented Reality, machine learning and brain computer interfaces. We help institutions to devise and implement their digital strategies, and to build the evidence base around new technologies

One thought on “Report: Building the Intelligent Campus – the Internet of Things

Leave a Reply

The following information is needed for us to identify you and display your comment. We’ll use it, as described in our standard privacy notice, to provide the service you’ve requested, as well as to identify problems or ways to make the service better. We’ll keep the information until we are told that you no longer want us to hold it.
Your email address will not be published. Required fields are marked *