Artificial Intelligence of Things (AIoT) is here: Is Canada ready?

The widely quoted phrase “the network is the computer”, coined in 1984 by Sun Microsystems VP John Gage, signified a trend that saw the transport of data become an essential pillar of information technology (IT). Today, computing devices that do not connect to the internet are virtually extinct.

In manufacturing and other fields, a similar fusion is taking place between cloud-based artificial intelligence (AI) technology and installations of Internet of Things (IoT) devices. The trend will see the operational technology (OT) that is used in industrial automation being enabled with the same computing power that is now being used to analyze medical images, predict disruptions in global supply chains and identify vulnerabilities in a corporate financial plan.

“AIoT is the fusion of two disruptive technologies, IoT and AI,” says René Breyel, AIoT strategic advisor at Montreal-based consulting firm Claridion. “IoT provides access to the state of any object or device in the physical world, and AI gives us the analysis of that data in order to understand and act on it.”

Connecting sensors to computing systems is nothing new – supervisory control and data acquisition (SCADA)–based monitoring systems date back to the 1960s. “We’ve had connectivity for oil rigs in the ocean for fifteen or twenty years,” says Breyel. “But those were multi-million-dollar solutions that were adopted because these rigs were really critical. Today, with the pricing drop of electronics and wireless connectivity, and with computing capacity and AI we get in the cloud, those amazing technologies are becoming widely available to every industry.”

Breyel sees three key areas where manufacturing will benefit from AIoT. “One key area is performance,” he says. “Because we get better information about machines and processes and their level of performance in real-time, we are able to process them more efficiently to make smarter decisions and produce more.”

The second benefit is better quality. “This is basically the same idea,” says Breyel. “You get more and richer data about what you produce, so you can control quality more effectively.”

Last but not least, AIoT can monitor the location of employees, forklifts, robots or machine tools to identify and mitigate safety risks. “Again, with more and better information, enriched by intelligent analytics, it becomes easier to understand and manage risk,” says Breyel.

Europe leads the charge

European companies, typically ahead of North America in industrial automation, are taking the lead in developing AIoT applications, and the most advanced use cases for the technology are in European countries.

Bosch operates a fully AIoT enabled semiconductor plant in Dresden, Germany. The IoT network monitors sensors, systems and products throughout the manufacturing processes, collecting the equivalent of 500 pages of text per second. AI-powered algorithms analyze the data in real time to predict trends and mitigate problems.

One of the advantages of this new capability is that the scope of data management can cover an entire manufacturing process from end to end. Waggeryd Cell Pulp Mill in southern Sweden, with production capacity of 175,000 tonnes per year, worked with Asea Brown Boveri (ABB) to deploy 156 smart sensors that monitor the entire power chain. This allows AI-enabled software to predict plant-wide trends based on power consumption, speed, vibration and temperatures.

To respond to this growing trend, Breyel joined with a few colleagues to found AIoT Canada, a not-for-profit cluster dedicated to educating Canadian business and government leaders about AIoT.

“It’s a story we started five years ago,” says Breyel. “We saw that we really needed an organization of people that are really involved with this to push the idea forward in Canada.”

At the outset, the organization turned to Europe for advice. “Bosch in Germany is a global leader in Industry 4.0 solutions and IoT connectivity,” says Breyel. “Four years ago, we brought them to Montreal to provide IoT training. They presented a digital innovation model but came to the conclusion that we didn’t have the maturity to push this kind of training at this level in North America at that time. For me, this was a wake-up call.”

It’s not just about technology

One of the big challenges is that AIoT involves connecting factory devices to the internet – something that was off limits in the past, for security and intellectual property concerns.

“There has traditionally been a gap between IT and operational technology (OT), the large industrial platforms that were implemented by companies like Siemens, Omron or ABB,” says Breyel. “These are used to manage all the different components in the factory. They have their own networks, their own protocols, their own security stack, and their own databases. But now we are coming up with solutions where IoT sensors are being connected to an IP wireless network going into the cloud. So now we want to bring those two networks together and implement IoT and AI stuff on top of that. This becomes quite challenging.”

One of the problems is that OT networks are completely new territory for IT people. “IT people are not used to managing OT systems because they were never allowed to interact with those operational and very sensitive infrastructures,” says Breyel. “The problem for them is that this factory equipment is in production 24/7, so as soon as you touch anything, it is risky and complicated. So, companies will need a strategy to deal with that.”

In general, it is the people issues that will be most challenging. “The technology is there, and people in manufacturing understand their processes and see the value in this, which is why AIoT is moving forward so quickly,” says Breyel. “The difficult and risky part is that the right people don’t have access to the training and education required to learn how to bring these two silos together. That will take a new degree of digital literacy across the organization and a digital transformation plan.”

Canada faces some particular challenges. One is that Canadian telcos favour cellular technologies and do not deploy LPWAN (Low Power, Wide Area Network), which is well suited to IoT networks and more widely available in Europe. Another is that relative to other countries, Canada has a large number of SME manufacturers that lack the resources to engage in a major technology transformation. AIoT Canada is building a community that will help mediate these issues. 

In spite of the challenges, AIoT Canada is seeing enormous interest from Canadian companies. The speed with which AIoT is developing gives industrial leaders little choice about getting on board with the technology. “The integration of these disruptive technologies within our factories is no longer an option,” says Breyel. “Industrial leaders must absolutely take them into account and integrate them into their strategic plan for their digital transformation.”

Jacob Stoller is a journalist and author who writes about Lean, information technology and finance.