Matt Johnson, Silicon Labs
The IoT, the IIoT, AI, smart buildings and smart cities are technologies that we are told will dominate our lives for the foreseeable future and make them markedly better.
That may or may not be, but one of the real issues for the end-user (and even the system designer) is how rapidly these technologies are developing and changing. Indeed, it is becoming increasingly difficult to keep up with the latest trends.
Take the IoT for example. It’s been said that it will take many years before consumers understand the full potential of being surrounded by networks of connected devices. Yet at the same time leading names in the sector are forging ahead with developing some of the most advanced and powerful IoT-related solutions.
Intelligence at the edge
One of the biggest growth areas in recent times has been the combination of AI and the IoT, which has been dubbed AIoT (the Artificial Internet of Things) and given no less a billing than the harbinger of the fourth industrial revolution (Industry 4.0). Specific sectors likely to benefit from AIoT include manufacturing (with ever-smarter supply chains), automotive (driverless vehicles), healthcare, and aerospace. One of the biggest advantages of bringing these technologies together is that it becomes possible not only to predict failure, as opposed to reacting to it, but also to correct problems automatically, referred to recently by Forbes as “self-healing”.
The equilibrium (some may say trade-off) between the two technologies is that while the IoT enables AI systems to interconnect, AI helps to turn data collected by the IoT into actionable information that facilitates the decision-making process. In other words, AI can make IoT data useful and form the backbone of any intelligent building or smart city.
A phrase that we will become more familiar with over the coming years is “artificial intelligence at the edge” or “edge-based AI”. This will see AI being used to enable edge site computing systems to handle increasing amounts of large data sets quickly and efficiently. Edge computing systems are used to overcome the inability of the cloud to handle large amounts of real-time data (which can also come at a hefty price). Then there’s the problem of device-to-cloud security, with data loss a common issue as well as intellectual property theft, online attacks, malware infections and general system vulnerability. Of course, there are other ways of getting around cloud security issues but they tend to be limiting and inconvenient – such as not using the cloud for sensitive data and carrying out regular back-ups.
Passing data across a network is always prone to risk but another emerging IoT trend designed to combat this is the adoption of advanced hardware-based security technologies. Ideal for complex supply chains connected via the internet and smart factories, these technologies have been designed to be affordable, scalable and highly secure; they are based around the use of security chips that offer protection against attacks and breaches.
Data security is also a concern in another IoT trend which has seen the increasing application of 5G wireless cellular technology in low-power wide area networks (LPWANs). While cellular systems are inherently insecure, 5G is known to be efficient and to offer high-strength connectivity, which is why it is finding increasing use in smart cities and IIoT applications.
Click image to enlarge
An alternative LPWAN-based technology is NarrowBand-Internet of Things (NB-IoT) which combines the advantages of low bandwidth with low-cost hardware and good coverage. NB-IoT also has limitations, however, because it is best suited to handling small amounts of data over short distances and, like 5G, is a cellular technology. One of the more interesting developments has been long-range (LoRa) wireless radio frequency technology which can deliver data reliably up to 30 miles away and offer extremely high levels of data security thanks to end-to-end AES128 data encryption.
Driving the IoT home market
The IoT has now progressed to a chicken and egg status – which came first, the network or the sensor, the various technologies or the demand for them? While we ponder this conundrum, what is clear is that driving the IoT market’s expansion has been the eagerness of consumers to control various aspects of their homes, whether through smart LED lighting, intelligent temperature control, smart security systems or automated and intuitive entertainment systems.
Click image to enlarge
The largest smart home segment of all the above is lighting and this is expected to send the market for wireless sensors stratospheric in the not too distant future. Energy management and security will be the next two biggest drivers of the increasing need for more sensor-based data. The initial slow take-up of smart home solutions will be just a vague memory as end-users become aware of what smart home ecosystems can deliver and as smart home networks get smarter.
Converging ecosystems and emerging architectures
Another trend that we will see developing is the convergence of the various ecosystems. This needs to happen because at the moment users have a potentially confusing choice of ecosystems, with Apple, Amazon Web Services (AWS), Microsoft and Google being the main contenders.
For example, Apple’s system is designed for its own proprietary devices operating at the edge while AWS connects physical devices to the cloud and manages data processing, storage, analysis and correlation to automate various tasks around the home (such as cleaning via the iRobot Roomba system). Google’s offering is based entirely around data in the cloud with devices at the edge while Microsoft aims to connect smart devices to its own smart services (plus non-Microsoft operating systems and devices). The so-called “intelligent cloud” includes various server products and cloud services such as Windows Server, SQL Server, Azure, System Center and Visual Studio.
At the same time, RISC-V processing architectures are emerging for wireless IoT devices. These ultra-low power solutions make it possible to manage major variations in IoT system workloads as well as timing constraints, voltage and other factors.
The future needn’t be confusing
We started by saying that it can be difficult to keep up with rapidly changing trends and each of the areas touched on above easily warrant their own lengthy individual focus. However, it’s also possible to become held back by the tendency to portray IoT adoption and integration as far more complex than it needs to be.
For the future, all you need to be concerned about is identifying your needs and choosing the best IoT specialists, such as Silicon Labs, to meet them. But if you take one key message from everything here it is that while the IoT in its various guises is indeed the future, like any enabling technology it needs to be easy for the end-user to implement and it needs to generate tangible benefits – otherwise, it just becomes technology for technology’s sake and nobody wants to invest in that.