Artificial intelligence is extending the potential for automation and paving the way for product-neutral production environments
Technologies centered on artificial intelligence (AI) are already enriching our daily lives, with examples including image and speech recognition. The technical background is complex: artificial intelligence is essentially based on the calculation of probabilities and the recognition of patterns. In an industrial setting, applications based on AI algorithms offer highly promising opportunities. This is because production requirements go way beyond today’s repetitive robotic applications.
In many areas of everyday life, AI has already resulted in spectacular solutions to problems that could not be solved easily, if at all, adopting conventional approaches. These include image and speech recognition or secure payment with credit cards. It is also foreseeable in robotics that new tasks can be automated with AI without having to be programmed for it. AI is far from the answer to every problem. The programming effort can be reduced, the operation becomes easier and processes will be more flexible. The decisive factor is the question of what should be produced. Product-neutral production cells, as already available in KUKA’s SmartProduction Center, are, as the name suggests, very flexible. This means that car doors can be produced today and washing machines tomorrow. In this way, production facilities can be adapted quickly and easily to new requirements.
When large data volumes reveal patterns
Production equipment and components are becoming increasingly networked: that is what Industrie 4.0 is all about. Artificial intelligence enables the effective utilization of the data thus acquired. For this, the corresponding applications require not only vast quantities of data, but above all the right data. In other words, Smart Data rather than Big Data.
The added value for companies is obvious: not only possessing data, but also being able to use them. To this end, the networked machines and robots send their data to a software or cloud application, for example. From the data volume, AI algorithms identify specific patterns and anomalies. General information about the production process is obtained in this way, for example information about day-to-day manufacturing sequences and forthcoming maintenance work. This “predictive maintenance” makes it possible to detect impending malfunctions in advance and thus prevent them from occurring in the first place.
The major challenge for intelligent machines is to solve tasks that are difficult to formulate as mathematical rules, such as speech recognition or the recognition and identification of images and faces.
How machine learning enables product-neutral manufacturing
Until now, robots have largely been predestined for repetitive applications. They perform their prescribed tasks with consistently high precision and repeatability. The production of the future involves increasingly complex requirements. One variety of AI – machine learning – is getting robot systems ready for flexible production. This involves interpreting data, finding correlations and deriving information from them.
In current production, sequences are extremely efficiently coordinated with one another; even brief interruptions or downtime have enormous economic consequences. Machine learning and AI have the potential to optimize productivity and availability during ongoing production. Further improvements are achieved in the areas of process quality, cycle times, energy consumption and maintenance intervals. This is enabled by means of central planning. A software package based on AI algorithms independently controls the production process. Not according to conventional concepts, however, by telling the machines how something is to be produced. Instead, the software plans what is to be done, taking cycle times and delivery times into consideration. For implementation, all that matters is what production resources are available.
Artificial intelligence encourages new forms of human-robot collaboration
In addition to these technological changes, there is also scope for entirely new forms of cooperation between humans and machines. Programming machines, or stopping them in an emergency, simply by talking to them is entirely conceivable. So, too, is dynamic image recognition for programming by demonstration. Further potential applications include intelligent service scenarios that can be displayed directly on the robot using augmented reality glasses. Or tablets that can be guided by a robot and display information. There are countless ways in which artificial intelligence can make our lives easier. AI is not just affecting our private lives, making it easier for us to handle machines via speech recognition or enabling us to make more secure payments, but is also transforming industry: versatile, optimized production that can respond quickly to individual requirements.