According to the Oxford Dictionary, artificial intelligence (AI) is the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. Already in 1956, AI was introduced as an academic discipline in a workshop at the Darmouth College (USA).

I wrote, as an editor, several articles about AI beginning of the 90ties. But then AI was for several years not more in the headlines of publications – with some exceptions, of course. Nowadays, AI is back: Everyone talks about ChatGPT and its competitors. ChatGPT stands for “Chatbot Generative Pre-trained Transformer”. It is an AI chatbot and thus a generative AI. OpenAI, the company behind ChatGPT, was originally founded by Elon Musk and Sam Altman, and in 2019, it secured a billion-dollar investment and cooperation from Microsoft. As a generative language and text-based chatbot, ChatGPT is capable of understanding written content – it can also compose texts itself. It works somewhat similar to interacting with a human. The AI understands the intention behind the written text, can engage in conversation with you, or even independently create texts and code.

But do not forget, AI does not lack shortcomings, which is why it still cannot replace a human, especially in decision-making. The overall data quality is rather subpar, lacking any verification mechanism to ensure the validity of used sources. Nevertheless, AI is already used in manufacturing operations: From quality management to product design. According to a study by Capgemini, more than 50 percent of top European manufacturers are currently implementing AI in some way. Germany leads, with 69 percent of manufacturers using it. In Japan that figure is 30 percent and in the United States it is 28 percent. China brings up the rear with 11 percent.

AI solutions need increasingly data to feed the AI algorithms. CANopen networks can provide this data. Sensors and measuring devices collect the data, CANopen host controllers pre-processes them, and standardized interfaces can be used to communicate with the AI engines. And vice versa, AI algorithms can improve the CANopen-connected actuators. Typical examples are AGVs (automated guided vehicles) and AMRs (autonomous mobile robots): AI can optimize AGV fleets regarding dispatching routes to maximize product throughput. Similar AI algorithms can be used to optimize throughput and energy consumption in elevators. Some of the needed input data is typically provided by CANopen devices and elevator cars and doors are actuated by other CANopen devices.

AI also can improve automatic welding solutions. Every welder knows how to adjust basic parameters such as amperage and line feed speeds according to different base materials, root gaps, and other variables. Nevertheless, welding is somehow unpredictable. AI-guided software can learn and teach the welding robot to do it better. This is even possible in manual welding, when the welder gets suggestions from a semi-automated equipment to do the welding in such a way that smoke is minimized. Recently, CiA has started to develop a CANopen profile for welding and cutting.

These are just some examples, in which “CANopen feeds AI”. Let’s talk about AI and CANopen in more detail during the SPS tradeshow in Nuremberg (November 14 to 16, 2023) at the CiA booth (hall 5, stand 410).