Artificial Intelligence (AI), and the Internet of Things (IoT) are increasingly becoming mainstream in the manufacturing industry. What is perhaps less understood is the degree to which these technologies are converging.
AI takes manufacturers to the next level by leveraging the data collected by the IIoT. Proactive intelligence, which is what we call it at Plataine, essentially means that the system analyses all variables and datasets to predict and solve problems that are about to occur. It does not wait for the user to ask for an analysis, it just does it, and then offers recommendations and insights to the factory floor operators, suggesting actions they can take to further optimize the process.
Given all this, it makes sense to begin thinking about the next concept – the AIoT – the Artificial Intelligence of Things.
Why talk about the AIoT?
This is a trend that is already out there. If the internet of things means connecting your ‘things’ to the internet, then AIoT means combining the power of AI with sensor driven data to make your manufacturing process intelligent. Let’s examine an advanced Aerospace & Defense manufacturer that is looking into adopting AI and IIoT technologies. What does this mean?
It means connecting manufacturing machinery (and, in addition, shop-floor staff, raw materials, tools and work-in-progress items) to the internet using sensors and machine connectivity. It means constantly collecting huge amounts of data in real-time, processing and cleansing it on edge devices, and transmitting all useful information to the cloud. But what comes next is different, because collecting and storing data is, of course, only the first part of what manufacturers need.
The next step is analyzing the data, and acting upon it, and it is here that we enter the realms of Artificial Intelligence and the AIoT. Sophisticated AI-based algorithms are required to understand how to interpret all that data and put it into a manufacturing context. Such algorithms can look at millions of datapoints streaming from a factory in real-time, as well as historical data, and apply artificial intelligence to both optimize operations for the long-term, and to solve problems in the short-term.
An example of optimizing operations for the short-term might be an AI based Digital Assistant that has realized a production deadline is not going to be met on time. In this scenario, the Digital Assistant analyses all variables and constraints to reschedule the production plan and ensure that things are bought back on track. It then alerts factory managers that there is a problem and recommends the new optimized plan. An example of AI solving long-term problems could be redeploying underused tools across a factory, to eliminate production bottlenecks and optimize their service cycles and OEE.
The most sophisticated AIoT solutions can also help to make data driven decisions– they don’t just tell management what is happening (insights) – they actually deliver recommendations. These intelligent insights can be fully autonomous – where the AI just does something to improve or help the factory and does not require any human intervention whatsoever; or they can be semi-autonomous, where the AI communicates to a factory employee to augment the employee’s knowledge and understanding of the situation: describing the problem and then giving them a recommendation how to solve it. And yet AI cannot function alone: it needs IIoT sensors deployed around the factory, recording all data in real-time and reporting back. Working in this way, AI and the IIoT come together to form the AIoT and becoming a true Digital Assistant to all factory staff.
Industrial IoT providers such as Plataine have huge experience with creating such AI-based solutions, based on their extensive work in the advanced manufacturing space. Plataine’s Digital Assistants deliver true proactive intelligence because they are aware of, and are always analyzing, all contexts and variables in the factory in real-time. For example, if incoming rush orders require immediate handling, an intelligent Digital Assistant can proactively modify the shift schedule or offer the shift manager an optimized schedule that takes work order priority into account or adds a third shift.
AIoT at the Edge
AIoT is deployed at two levels: at the Edge, and in the cloud. At the edge level, it makes sense to do as much filtering and cleansing of the data as possible before transmitting it to the cloud. Ideally, a manufacturer does not want to be sending every bit of data to the cloud. Doing this is expensive, both in terms of data transmission and in terms of storage; moreover, it increases latency – and therefore potentially introduces unnecessary inefficiencies on the shop floor.
AIoT in the cloud
AIoT at the Edge will only ever be part of the story. To get the highest-level of benefits that AIoT technology has to offer and to get the most out of AIoT in terms of actually turning your data into decisions – it’s necessary to do most of the heavy analysis and processing work in the cloud. An AIoT solution – even a solution that has been deployed only in a single factory location – is likely to depend on multiple edge devices to take the data out of the factory, compile it, and transmit it to the cloud. Until all that data has been put together in one location in the cloud can form a complete picture of the factory, the benefits AI can deliver are limited.
Additionally, many AIoT solutions are not limited to a single factory but can be based across several factory sites. Multinational firms may be running an AIoT solution across factories on different continents. Manufacturers may also use AIoT to connect to their suppliers to optimize the supply chain – monitoring raw materials to ensure they are being produced according to schedule without compromising the quality levels required and are being transported in a safe manner that guarantees they arrive at the factory on time. Or they may be using the technology to connect with customers – so that they can better assist with post-sales service and support.
With millions of datapoints available, a manufacturer will only get the true value if all that data is considered as a whole, so that the AIoT solution can apply analytics and machine learning to identify underlying trends and patterns that have never previously been visible.
It won’t end there…
AIoT is a fast-moving technology and, ultimately, the smart factory will require that AI-based solutions will need intelligence deployed at every possible level. It is likely that AI capabilities will provide manufacturers with additional capabilities – as they accumulate more historical data, machine learning algorithms will continue improving the predictions and recommendations and the decision making will be easier and optimized.
To continue the conversation about the AIoT, get in touch with your local Plataine Digital Transformation Specialist today.