Getty Images/iStockphoto

Guest Post

Scaling Industry 4.0 with AI, IoT sensors and hybrid cloud

To get the most value from any industrial process, organizations must have the technology to collect IoT data, analyze that data with AI and do so in real time.

AI's growing sophistication means that organizations of any size can much more easily use AI to solve critical, complex problems.

AI has been pivotal in navigating the particularly challenging landscape of the past year. Retailers have relied on AI to help them optimize their order shipments, reimagine their stores as distribution centers and ensure people can still get products even when in-person shopping ground to a halt. In the utility sector, AI has been increasingly deployed to keep power grids running by managing issues such as vegetation risk or getting an earlier start preparing for adverse weather events. AI is also helping build better from the ground up. For example, AI analyzes and tracks requirements for complex engineering equipment used in can't-fail scenarios such as airplanes, ventilators and space shuttles.

When AI is combined with other enabling technologies, we begin to see some of the ramifications of so-called Industry 4.0. When combined with IoT, AI can analyze sensor data and predict the failures of industrial assets like factory equipment, HVAC systems and assembly lines. It can optimize the schedule of asset work orders, analyze the risk of failure and let managers prioritize repairs under different criteria. Visual inspection is being used to detect manufacturing defects and help keep workers safe by analyzing real-time video feeds.

Cameras, beacons and sensors can monitor a facility 24 hours a day, seven days a week. With the help of AI that can separate signal from noise, organizations can ensure no valuable insights are missed, and begin to automate increasingly complex parts of their manufacturing and production process. These building blocks of Industry 4.0 are already mature and enterprise-ready if companies invest in the needed underlying digital infrastructure.

Unlocking Industry 4.0 with hybrid cloud

AI and IoT are two of the key building blocks toward large-scale industrial automation, which is what we usually mean when we talk about Industry 4.0. However, achieving any of the above applications at scale also introduces new challenges that require a third building block: hybrid cloud.

Think about the amount of data inputs on a single factory floor, from IoT sensors that track heat and occupancy to the cameras collecting visual data and monitoring workplace safety. Extrapolate that across a larger organization with several different facilities, perhaps even different kinds of facilities, and the amount of data to be processed grows exponentially. The AI models needed to sort through all the data get significantly more complex. And perhaps most importantly, time becomes an issue. A model that can tell you a month from now that employees are crowding in a certain passageway isn't particularly useful. To take advantage of predictive insights, there needs to be the ability to act on those insights right away, which means being able to compute at the edge where these insights are gathered.

These three components -- the ability to gather and store vast and changing amounts of data, the ability to run models or other software on top of that data, and to be able to do it anywhere you want -- require an infrastructure footprint that stretches from the edge to the data center and to the cloud.

To be efficient, organizations need a seamless management plane across all infrastructure. Hybrid cloud facilitates this with a common container-based platform that can run across all infrastructure locations. It gives the ability to autoscale based on workloads, and the flexibility to run a platform on any cloud, public, private or at the edge.

In an Industry 4.0 context, hybrid cloud can connect the dots. It makes the data, AI, tools and software that employees need available wherever they need it to be. The easier you can make people's jobs, the more time, attention and ability they can devote to solving even more interesting, complex and expensive problems.

About the author

Rishi Vaish is the CTO at IBM AI Applications, a portfolio of enterprise applications spanning asset management, facilities management, supply chain management, engineering lifecycle management, weather business solutions and blockchain. Through his career of more than 15 years, he has developed a deep expertise in AI-based applications, hybrid cloud technologies, cloud computing, SaaS and application middleware. He is passionate about driving innovation, modernization and transforming product and technology organizations to meet scale and growth demands.

Dig Deeper on Enterprise internet of things

Data Center
Data Management