AI & IoT

A Perfect Match: AI & IoT

It is supposed to change everything: artificial intelligence. According to a Bitkom study, 60% of the companies surveyed are convinced that AI is the most important future technology. At the same time, experts now agree that AI and IoT should not be considered separately. Why is that?

Through the Internet of Things, enormous amounts of operational and production data are collected and evaluated in many areas. In order to generate real added value from this IoT data, companies need AI and analytics tools to help quickly identify problem areas (anomaly detection), derive decisions and automate optimization measures. Only the combination of AI & IoT can bring that decisive breakthrough in digital transformation.

AI & IoT in practical application

When it comes to the practical application of AI & IoT, some companies are already ahead of their competitors, as shown in the Bitkom study 2019. Swabian technology enterprise Voith uses KI & IoT in the service and maintenance of large facilities. Important factors here include a metadata model that brings together all the information an engineer needs to commission a machine as well as the use of machine learning to analyze raw data on which basis smart services can then be implemented.

Starting with simple ML algorithms

All the hype about AI leads many companies to imagine the application of AI and machine learning to be highly complex. In fact, practice shows that simple ML algorithms, such as linear regression, often have a significantly stronger prognostic power than complex ones, such as Gaussian distribution or k-Means. “In many use cases, linear regression is sufficient as a statistical analysis method to make predictions about when machine parts or tools need to be replaced”, says Hendrik Nieweg, VP Solution Management at Device Insight.

For example, Device Insight implemented an algorithm for robot manufacturer KUKA that predicts when the next maintenance is due for a specific robot type. The forecast is based on usage data from the robot and uses linear regression as well as a “Generalized Additive Model” (GAM)-based forecasting method. The advantage of this is that the robot is only maintained as required, rather than too often or unnecessarily. From a company point of view, such predictive maintenance naturally makes much more sense.

More efficiency, more sales – more satisfied customers

Many companies are still in the early stages of developing an IoT application that gives them a transparent insight into their operating and production processes. However, the next stage – increasing reliability, efficiency, and productivity – is already imminent. This is where sophisticated AI functionalities come into play, especially ML algorithms.

Used together, AI & IoT open up numerous opportunities for companies: amounts of complex data generated by IoT projects can be evaluated by AI at high speed; knowledge gained flows directly into improving production processes and product quality, which in turn leads to more efficiency in production, more sales and returns and, last but not least, to higher customer satisfaction.

Latest Posts



X
Tweet
Share
Pin
Share