The IoT - what’s so special about it?

The Internet of Things  (IoT) is a confederation of technologies that allows industrial and public enterprises to gain deeper insight into the processes they’re running, and improve operational methods and efficiency accordingly. It’s also making life better and more comfortable at a personal and domestic level. It’s achieving these ambitious objectives by using large quantities of sensor and actuator edge devices that communicate across the Internet with powerful yet affordable cloud-based data processing and analysis resources.

But how is the IoT different from everything that’s preceded it? One way to answer this is to look at some successful project examples, and define the IoT in terms of how it’s benefiting its users. We can then complement this by reviewing the building blocks of an IoT infrastructure, and seeing how they work together.

Both approaches are included in our discussion below. First is a description of the IoT as a ‘control system plus’, which introduces the extent of its power and potential. We show how the concepts of ‘IoT’ and ‘M2M’ relate to one another. We fill out this picture with a set of examples that demonstrate various aspects of the IoT’s contribution to the world around us. These descriptions are complemented by a summary of the IoT’s key building blocks. Finally, we suggest a couple of ways in which developers can start from basics, and find out for themselves what the IoT could mean for them; the opportunities that await, and some practical tools for achieving them.

The IoT - a control system and so much more

Not everyone has heard of the IoT, but the concept of ‘smart devices’ such as intelligent fridges or remotely-controlled HVAC systems is better - known. Such devices can be components of an IoT system, but they’re only a part of the entire concept. A better way to appreciate what the IoT truly is, and of its immense potential, is to look at it as a ‘control system plus’.

A traditional control system – as found throughout industry for many decades and still very much in use today – might be used to manage a production tank in a processing plant. The system could comprise tank temperature and level sensors, a programmable controller and user interface, and a heating jacket and agitator. An operator tells the process to start; the controller responds by checking the level sensor to ensure that there’s product in the tank, and then energises the heating jacket until the temperature sensor indicates that the required temperature has been reached. The system also runs the agitator throughout the heating process.

Implementations like this, based on the core concept of sensor-controller-actuator, and known as closed-loop control systems, work well, and are certainly more efficient and reliable than asking a human operator to manually control large numbers of tanks. Yet think about how much more such systems could contribute to their owners’ efficiency and productivity if only they had access to more data, and were better joined-up to other resources around the factory.

For example, if the control system had information from the company’s sales office about order status and stock levels, it could decide on whether to run the process now, or defer it; and a centralised system could queue the tanks optimally, so oldest production batches are sent to the packaging area first. Additionally, a sensor on the agitator could, over time, detect changes in vibration levels; the controller unit could then recommend preventative maintenance action and order spare parts before the agitator fails.

While this is a description of one factory, current and future IoT implementations typically benefit multiple manufacturing and corporate sites over a large geographical area. Unexpected demand in one plant, for example, could be mitigated by a central system that could detect spare capacity in another plant and switch production accordingly.

Ideas such as these – richer data, more connectivity and greater insight – are the driving concepts behind IoT projects as they appear, not only in factories and homes, but also city infrastructures, emergency services, transportation systems and highways management, retail chains, personal wellbeing, patient monitoring and wearable devices, and anywhere else where innovators can apply IoT solutions to improve living standards or boost efficiency and productivity.

Although widely diverse, the IoT implementations that service them share a set of common elements that can be defined, and which characterise the IoT itself. We will look more closely at these building blocks later, but, in summary, the IoT’s success depends on sensors and edge devices that are low enough in cost and power demand to allow deployment in volume, local area networks and gateways that can collect and process edge device data intelligently, and scalable, cost-effective yet powerful cloud-based computer resources to handle enterprise-wide data processing, analysis and storage.

This combination of rich data, often from apparently unrelated sources, and sophisticated computing and analysis power provides deeper than previously available insights into the world about us, so systems and humans can make better-informed and more effective decisions. IoT applications for consumers work with household appliances and other consumer devices to help people live better. By contrast, ‘Industrial’ IoT as Microsoft  refers to it – concerns organisations’ use of intelligent devices and the data they collect to streamline and improve operations.

IoT and M2M networks

M2M simply means ‘machine to machine’ and refers to machines and devices that are connected to one another, and to the Internet. The distinction between M2M and the IoT is blurred; sometimes they may be the same thing. In other cases, the IoT may be providing a bigger picture; generating actions and recommendations from a wider range of inputs than the underlying M2M network. Alternatively, an IoT implementation could comprise several M2M networks.

Imagine, for example, a road junction managed by a set of traffic lights. Sensors could monitor the traffic flow in each road approaching the junction, so the signal phasing could be modified accordingly. The junction would be managed safely, but in isolation. If, instead, the traffic lights’ M2M control system data – alongside that from all the others in the area – were fed into a citywide IoT traffic management system, traffic flow through the entire area could be optimised from an all-seeing central resource.

Some examples of the IoT in action

The Things Network  provides a simple ‘smart mousetrap’ example that clearly illustrates both an IoT strength and a constraint.

For a facilities manager responsible for a large industrial, commercial or residential area, the need to regularly check large numbers of mousetraps around the premises presents a significant workload. Instead, if each mousetrap has a sensor that can signal back to an application when a mouse is caught, staff only need to visit ‘triggered’ traps rather than the full complement, considerably reducing workload.

However, there is a caveat; each trap sensor must be so light on power demand that it can operate by energy harvesting, or at least with a battery that will last for several years. Otherwise, time and money saved by eliminating inspection of empty traps will be lost again in replacing spent batteries.

The next three applications look at a medical implant, smart home, and city infrastructure application respectively, to demonstrate the scalability and diversity of IoT technology.

-Medical implant

In January 2017, IoT Evolution magazine ran an article  about the role of the IoT in providing chronic pain relief. It describes how chronic pain therapy is provided by implanted devices that send small, highly-tailored electrical pulses directly into the spinal column. Currently, these devices must be updated periodically to restore their pain relief capability. Without network connectivity, this can mean inconvenience and even suffering, as the patient must travel to the doctor to allow direct coupling to the device.

Accordingly, next-generation devices will have Bluetooth communications, so updates can be received and usage data sent via smartphones and the Internet. This will eliminate the need to travel, while daily collection of usage data will improve pain management offerings and allow plans tailor-made to the patient’s needs.

Improved, remote control of pain-relief implants like this is just one example of how the IOT is impacting medical applications. Increasingly, medical centres will be able to remotely monitor large numbers of patients, not only for pain relief but also for cardiovascular diseases and many other conditions. Additionally, some systems, while monitoring conditions, are also using the same communications channels to administer injections or take other action – without involving the patient or the doctor. Freeing medical staff from these essential yet time-consuming activities is a major productivity boost, while patients enjoy a less-inhibited lifestyle.

-Smart homes

While many consumers are now aware of smart devices, and some own them, the true potential of a smart home remains unfulfilled. Intel  believes this is partly due to the high level of fragmentation in the industry, leading to countless offerings that don’t work together.

The company’s response is based on its product breadth and ecosystem scale, as exemplified by their smart ‘Tiny house’. At just 210 square feet, this demonstrates what is already possible and explores what is needed to take the home from ‘connected’ to ‘truly smart’, and become widely adopted.

Connectivity must be made simple, industry standards adopted and security ensured across the physical and digital worlds, while data between devices and between neighbourhoods must lead to insightful actions. Cloud connectivity, advanced device management and built-in security will connect consumers to a variety of new services, features and cost savings.

-City Infrastructure

CityVerve  is an innovative IoT project located in Manchester that has won a £10 million-pound Government-led technology competition. The project, led by Greater Manchester Local Enterprise Partnership, was selected because of its ambition, scale, coordination across the public and private sector, and potential for success. As well as bringing real benefits to people who live and work across Manchester, the project will help the UK to be a world leader in the adoption of Internet of Things technologies and inspire others around the world to create smarter cities.

The project comprises several innovative elements, including sensor networks in public places that promote people’s wellbeing, health and safe enjoyment of outdoor activities. These are complemented by ‘talkative bus stops’, smart lighting and crowd-sourced bike sharing services that offer safe and attractive alternatives to excessive dependence on cars.

IoT building blocks

The applications discussed above, while highly diverse in scale and application, are based on a common set of IoT ‘building blocks’. Within each building block will be a mix of hardware and software elements highly specific to each application, yet the blocks will be engineered to work with one another, from end to end of the IoT implementation, in accordance with a common model recognisable to all IoT designers.

Fig. 1: Generic IoT model

The lowest level of Fig 1’s model is at the IoT’s edge, and comprises edge devices containing sensors or actuators. Each edge device will also include a power supply – usually a battery or energy-harvesting system - and an embedded microprocessor and communications interface. The microprocessor collects the sensor’s real-world measured data, packages it and transmits it through the communications interface onto the chosen local network; Wi-Fi, Bluetooth, ZigBee or one of many others. The choice is a trade-off between transmission distance, security, reliability, data rate and power consumption.

For developers, finding an edge device solution is often easier than the above paragraph implies. As suppliers compete for market share, they are introducing smart products that contain all necessary sensor (or actuator), processing, security, power and communications functions within one small, attractively-priced, off-the-shelf unit. This can add up to a lot of functionality; process sensor functions can include not only the sensing element, a-d conversion, and communications interface, but also self-testing, self-validation, self-calibration, self-diagnosis and signal processing.

Smart actuators are also available. The integrated communications and processing functions that deliver actuation commands can also carry feedback signals relating to position and speed, as well as from temperature, load or other sensors. This feedback can be used to synchronise the activities of a large number of actuators, as well as monitoring their health and spotting potential problems.

Also on the network is a gateway, which aggregates all the sensors’ data, and to some extent processes it; possibly driving a local user interface display, or sending commands back to edge actuator devices. The gateway should also handle security issues such as ensuring that only authorised ‘white-listed’ code runs on the platform, while unauthorised changes cannot be made. Access permissions to measured data must also be managed.

The gateway will package measured and calculated data for transmission across the Internet via one of several channels. If it’s in a remote location without an existing communications infrastructure, it will use a cellular wireless standard such as NB-IoT, or possibly a non-cellular, low-power wide-area (LPWA) network such as SigFox or LoRa. In other situations, it will connect via Wi-Fi or an Ethernet cable link to a router that handles onward data exchange.

The gateway’s Internet connectivity provides access to a centralised Cloud resource that can perform further processing, integration and correlation of results with data arriving from other gateways. From this vantage point, and with sophisticated analysis tools, the cloud resource can see the bigger picture and offer recommendations or control actions based on a deep insight into the entire process.

Crucially, the powerful computing tools essential to delivering these insightful results are available to organisations of all sizes at affordable prices by using subscription-based services such as Software as a Service (SaaS). Users simply access the resources when they need them rather than funding a data centre or IT room with all its hardware, software, environment, security and staffing issues.

Against this generic reference, a single road junction in our traffic light example would look like Fig. 2.

Fig. 2: IoT Traffic light example – single junction

Maintenance-free magnetic vehicle detectors are available which can be buried in the road, have an IP 68 protection rating, transmit data and receive update information via a ZigBee wireless network, and have a battery specified for 5 – 10 years’ operation. They correspond with a wireless access point, which acts as a gateway between the wireless sensor network and the Internet. Besides exchanging data with the centralised resource, the gateway manages the traffic lights using both centralised data and feedback from the local vehicle sensors.

Several such systems working together can monitor traffic flow over a wide area and implement dynamically adaptive control of the area’s traffic lights. Traffic flow is improved while CO2 emissions are reduced. Fig. 3 shows a model of a complete system where multiple road junctions are brought into the single IoT implementation.

Fig. 3: IoT wide area traffic management scheme with multiple road junctions

Your IoT

If you’d like to move from a conceptual consideration of the IoT to a practical evaluation of what it could do for your organisation, here are a couple of tools that could get you started:

o Microsoft Azure Events Hub:  Connect an Arduino or Raspberry Pi board to some off-the-shelf sensors (temperature, light, motion, sound, etc) - then transmit the measured data to the cloud and run sophisticated data analytics on it using a set of Microsoft Azure services.

o Join The Things Network  to take part in building an IoT data network with a global community of 13861 people in over 84 countries. Learn through experimentation with low-cost devices, and by forum discussions with fellow members.

Conclusions

We have seen that ‘smart devices’, ‘connectivity’ and ‘network’ are all key words related to the IoT, but none encompass it completely. ‘Control system’ is closer, as the IoT operates all the control system components; sensors, controller and actuators. The difference between the IoT and more traditional control systems lies in the far-reaching, rich and ubiquitous distribution of smart edge devices, a wide choice of local area networks, and the availability of cost-effective but powerful cloud-based computing services.

Developing solutions from these elements is not without challenges, as many different suppliers, technologies and standards are competing for market share, and interoperability is not a given. For example, gateways must communicate with edge devices, and off-the-shelf cloud platform solutions are unlikely. Some manufacturers like Intel claim sufficient breadth of technology to offer complete, end-to-end solutions, but these will not cover all applications.

Yet, as this article has also shown, these challenges are being overcome, to fulfil projects of all sizes from personal to city-wide. They are realising the IoT ambition of bringing deeper insight and better control of the world around us than ever previously available.