The confluence of Artificial Intelligence and Industrial Internet of Things may be seen as the sunrise on a new digital revolution for manufacturing organisations and the birth of Internet of Things Data as a Service (IoTDaaS).
Artificial Intelligence has been a topic for many news articles and blogs over the past few years. The term AI is fluid in its definition to as many people as there are articles, a bit like defining what "Cloud" was when it first emerged. One of the technology trends emerging in 2022 is the Artificial Intelligence of Things, (AIoT), which is hoped to redefine the future of industrial automation and set to lead the Industry 4.0 revolution. Indeed, Industry 4.0 is dependent on the processing power of the previous Cloud revolution. Cloud technology gave us the ability to aggregate IoT data into Big Data architecture systems such as Hadoop and Spark and crunch it into meaningful charts to give insights for stakeholders. AI takes this to another level by providing the actions based on the data it collects and analyses, rather than waiting for operators to initiate responses.
The basis for the Industrial Internet of Things (IIoT) is the ability to collect massive quantities of data and make these integrated datasets accessible across the organization for strategic decision-making. Historically, industrial manufacturing has not typically had the strategies or focus to provide their own capabilities in data science, which includes artificial intelligence and machine learning. However, such skill sets and environments are now readily available for manufacturing to take advantage of the plethora of … as-a-service cloud services testify to this.
There is a need to extend the intelligence out from the cloud centre to the edge collection points. The transmission of data from IoT devices into the cloud can be both overwhelming, from a volume perspective as well as expensive when the network for devices is exclusively cellular. Even if the network transmission is relatively cheap, the cost of cloud IOPS to filter out the required data from the chaff of polled data is unnecessary if the data is filtered. Edge intelligence to provide a logical filtering and event correlation is the obvious answer. Whilst the usual edge processing cannot be powerful enough for true AI it can certainly provide the kind of intelligent logic to provide the filtering and event correlation services.
At Simoco we natively provide these types of intelligent devices across many diverse industry sectors. The aggregation of data allows for more meaningful data be presented as a holistic service. Many companies have had to cobble disparate IoT feeds into a central system, this has led to large volumes of data being transmitted and filtered within the cloud infrastructure where the cost of cloud IOPS and cellular data usage have pushed the cost way up.
The cost of data transmission over disparate multiple SIM contracts and the cost of filtering out spurious data from the required nuggets is a financial burden that may inhibit the agility a business needs. It seems a reasonable prerequisite to filter and transmit only the data that is required and companies that can provide an end-to-end service on IoT data are set to rise.