first ifsecond if
AI and Machine Learning

AIoT benefits for manufacturers

The Artificial Intellience of Things (AIoT) can offer many potential benefits for edge and cloud computing applications for manufacturers.

By Control Engineering Europe February 27, 2021
Courtesy: CFE Media and Technology

The Artificial Intelligence of Things (AIoT) refers to the adoption of AI capabilities in Internet of Things (IoT) applications. In simple terms, while the Internet of Things (IoT) is reactive while Artificial Intelligence of Things (AIoT) is proactive. The AIoT relates to the combination of AI and the IoT with the goal of achieving greater efficiency in complex operations – it looks to improve human and machine interactions and to enhance data driven decision making. AIoT also adds the ‘brain’ to the Cloud and to the Edge.

What effect do you think AIoT will have within the industrial sector?

Giuseppe Surace, chief product and marketing officer at Eurotech (GS): AI technologies help companies to leverage the potential of the huge amount of data generated by devices, assets or ‘things’. Immediately sending all this data to the cloud or a data center would require an adequate infrastructure, which would be expensive to deploy and maintain, without taking bandwidth and latency issues into account. Bringing AI, machine and deep learning closer to the source of data can help businesses to provide a meaning to the data collected in the field to turn it into actionable business decisions.

James H. Chappell, global head of AI and advanced analytics at AVEVA (JC): Fundamentally, as companies adopt AIoT they will become more efficient, reduce costs, and improve overall operations including accuracy and safety – and by reducing wastage and reducing their carbon emission during these complex operations they will become more sustainable. AI uses a wide variety of historical data to analyze trends, and this can help streamline and improve processes.

Pete Sopcik, ADI OtoSense marketing manager at Analog Devices (PS): AIoT is having a significant impact on several key areas in the industrial ecosystem – faster decision making; reduced infrastructure costs; and improved security architectures.  Each of these benefits are driven by the ability of AIoT to deliver insights at the edge.

In what sort of time-frame do you expect to see AIoT technologies being adopted in the industrial sector? 

Surace: AI is already moving to the IoT edge, thanks to the advances in edge computing technologies. By connected field devices to powerful edge computers, there is no need to send all the data to the cloud or data center, as it can be processed and analyzed on site. Although we are still far away from seeing completely autonomous applications (like level 5 autonomous driving), there will be a consistent leap in the amount of data processed at the edge over the next 5-10 years.

Chappell: AIoT is already being used in the industrial sector, but it is still in its infancy. Predictive analytics uses supervised machine learning to understand an asset’s individual operating history and develops a series of normal operational profiles for each specific piece of equipment.  When this is applied to the cloud it offers real-time operating data to detect the subtle changes in system behavior that are often the early warning signs of pending equipment failure. And it’s becoming more sophisticated by combining together cognitive technologies.  For example, machine learning applied to various IoT data can detect when heavy machinery on a manufacturing line is not operating at its peak.

Sopcik: We are already witnessing AIoT technologies adoption in industrial environments.  These are coming in the form of quality control applications such as monitoring end of line quality on finished goods.  Customers are creating models of their assets coming off the production line and testing these assets to validate operation prior to shipping to customers. Such applications are paving the way for broader adoption of AIoT solutions that are more tightly integrated with the manufacturing ecosystem, optimizing maintenance strategies, process control and manufacturing operations.

How do you expect AIoT affect the control engineering function?

Surace: Even though most of AI training will still happen in the cloud, there will be the need to deploy trained models to optimize field applications, such as predictive maintenance and robotics. The challenge is providing AI models with the right data. As AI training and inference happens at a software level, the role of control engineers will be crucial to provide that software with quality data to improve the machine and deep learning models.

Chappell: In order to get the most value from AIoT, it is important to understand what it can do today, and it is equally important to understand what it cannot provide. By matching business use cases to AI capabilities, companies can pragmatically leverage the power of AI in order to achieve maximum benefit with realistic and beneficial results. For example, a company having issues with unplanned operational downtime can leverage AI to provide early detection of issues based on the data provided by sensors they may already have in place.  Furthermore, they can leverage techniques such as deep learning to provide forecasts of an asset’s remaining useful life in order to better assess situational risk. By following best practices and integrating AI intelligence into business operations – through the Cloud – companies can maximize the benefits and value of their investments.

Sopcik: AIoT will play an important role in how control engineering impacts the industrial environment and adapts to new technologies that drive expanding capabilities.  AIoT will drive tighter integration between assets and process optimization. Faster decision times lead to better control.  Analytics on each product – not just sample sizes – will deliver better insights and improved accuracies of the manufacturing and quality control processes that govern the industrial environment.

Control engineers will need to continue the integration of these connected devices and improved information into the industrial ecosystem while also playing a key role to ensure that these analytics are appropriately applied and controlled to enable optimized, reliable and safe manufacturing conditions. Best practices and how to integrate these solutions to deliver optimal efficiencies must be considered to ensure this integration is effective and scalable.

Can you offer any ideas for practical AIoT applications in the industrial sector?

Surace: Predictive maintenance operations will benefit hugely from AI and AIoT. By collecting more and more data over time, maintenance processes will become increasingly efficient, leading to a consistent reduction of costs, with less waste and reduced downtimes. The more data that is used to feed a training model, the more efficient that model will become.

Another field of applications of AIoT technology is robotics and machine vision. Robots and cobots make large use of camera inputs to perform their task. The more data can be processed and analyzed in real time, the more their work will be efficient. This would enhance the quality of production processes, leading to better outcome, again with less waste.

Chappell: In order to get the most value from AIoT, companies need three primary things – a corporate culture that encourages and facilitates AI infusion into business processes; an understanding and trust of the power of AI capabilities; and an IT infrastructure that provides the underlying data requirements and processing power needed by AI – and that includes utilizing a Cloud Platform. Without all three, barriers will inherently exist to achieving maximum success.

Sopcik: Quality control is one application that is being addressed today in various forms. Whether it is monitoring pumps, compressors, sub-assemblies or final products such as refrigerators, AIoT is enabling new insights into the machine operation to address critical quality problems.

Bringing diagnostics directly to the critical equipment in the field is another key application for AIoT.  Large assets such as aircraft engines, wind turbines and vehicles require service teams to go to the equipment in the field. AIoT solutions can deliver analytics at the asset in the field, leading to corrective actions that can be taken immediately.

This article originally appeared on Control Engineering Europe’s website.


Control Engineering Europe