Internet of Things (IoT) is advancing rapidly, thanks to a flood of low cost, low power devices in the market.
Artificial Intelligence (AI) enabled command and control is the next milestone, IoT experts told delegates at an Institute of Information Technology Professionals South Africa (IITPSA) Special Interest Group in AI and Robotics (SIGAIR) briefing recently in Johannesburg.
Noting that the rise of the ‘Artificial Intelligence of the Internet of Things’ posed significant new business opportunities, speakers assessed the impact and role of AI-enabled IoT in future.
Craig Boshoff, COO of Real Telematics, highlighted the evolution of IoT from the early days of telemetry. “One drawback of wanting to monitor everything is the amount of data involved,” he noted. Not only had telemetry systems transmitted huge amounts of data constantly, but they had also been costly and consumed a great deal of power.
AI-enabled IoT now brings to market low cost, low power devices, which could transmit only exceptions, so reducing the amount of data needed. “Now, these parameters are often programmed into the device, but in future, machine learning will be used to identify normal parameters and allow for only exceptions to be transmitted,” he said.
Emerging from this, AI would start to act on exceptions, without the need for human intervention, speakers said. “The ultimate is ‘command and control’, in which a system sees an exception and does something about it,” said Pieter Pienaar, CEO of Informed Decisions and overall winner of 2019 MTN IoT Awards.
Internet of Things devices are of huge value
Pienaar said that while users would not want to have constant updates arriving on their phones about a system performing normally, the underlying data gathered by AI-enabled IoT devices was of huge value. “In my world, data is gold,” he said. “The disruption happening today is that the price of obtaining data is relatively low – it can be sent over long distances, using very little power. With data from IoT systems, overlaid with data from other sources, it becomes possible to identify trends in seemingly insignificant data, Pienaar said.
He cited an example of a manufacturing firm, which had reduced its reject rates by 45% after tracking humidity sensor data and quantifying the effect humidity had on the quality of its products. While the technology had progressed, IoT adoption was still in its infancy, he said.
“All of us in this industry are going for a land grab at the moment,” said Pienaar. “There are opportunities for businesses across the stack, and a significant shortage of AI and machine learning skills at the moment.”
“The possibilities are endless for business,” said Gerhard van Wyk, Associate Partner at IQbusiness. “We’re now in a position to look at how we can use these technologies to do things differently.” Van Wyk outlined his journey in fully automating his home environment using integrated IoT sensors. Although he said he had done so ‘just because I can’, he noted that he had achieved significant savings on water and electricity as a result.
“It’s not just about installing devices, it’s about what you do with the data from those devices,” he said. “But businesses cannot roll out widespread monitoring ‘just because they can’. They have to be pragmatic, and ask what benefits they can achieve from monitoring.”
Johan Steyn, Chair of the IITPSA SIGAIR and Senior Manager at IQbusiness, noted that AI and IoT were not just about devices and technology, however. “There has to be a business case. You can place devices everywhere, but business must know what it wants to do with AI and IoT.”