"When you think of robots, you might first think of the devices in your own home for vacuuming or mowing lawns. Or the classic large industrial robots from the automotive industry that lift car bodies or set welding spots.
The humanoid-designed robots such as Pepper or Tesla Optimus are comparatively new. But beyond that, the connection between robotics and artificial intelligence creates far more areas of application, for example in intralogistics, the service sector, agriculture, the healthcare industry and, last but not least, of course in various production contexts.
Numerous industries rely on robotics
The drivers for more automation are diverse. Above all, there are overarching developments such as demographic change, the shortage of skilled workers and increasing cost pressure everywhere, especially in a high-wage country like Germany. There are also industry-specific challenges. For example, the rise of online trading is driving intralogistics to become more efficient in its processes. And the ban on glyphosate reinforces the need for automated mechanical solutions for weed removal in agriculture.
While industrial robots operate in the production environment, service robots can be found in many other places, for example in healthcare facilities, hotels, restaurants, at train stations or in fields. The markets also differ: only various components such as controls, sensors and especially the tools make the industrial robot complete and profitable. Accordingly, there is an entire market of suppliers and integrators around industrial robots who turn the robot into a functioning robot system. In contrast, service robotics is much more product-oriented: Users, whether for commercial or private purposes, usually buy a complete system that carries out a specific task such as floor cleaning.
The so-called cobots, which are often used for welding applications or in autonomous mobile robots, are currently experiencing a boom. Robot-based gripping and packaging of goods in intralogistics is on the verge of a breakthrough. Other applications such as robotic assistance systems for care, however, require even greater research and integration.
Innovation boost through artificial intelligence
Artificial intelligence will help industrial and service robots achieve a breakthrough in many places - or at least get closer to it. We are currently experiencing two major trends that are being intensively researched. Firstly, there is cognitive, i.e. perceptive, robotics, which uses sensors to collect environmental data, processes it and derives appropriate actions from it. In the future, robots will often no longer have to be programmed down to the last detail for every application. This is particularly important in order to economically implement increasingly personalized production with small batch sizes and in this way to bring robotics into medium-sized companies.
Secondly, the current research is about “automating automation”. AI is already in use today in controlling robots or in image processing. In the future, it will also reduce the effort for engineering and commissioning by, for example, partially automating risk assessments or using simulations to train the robot to suit the task to be carried out. If you look at the number of new installations of industrial robots every year - around 550,000 worldwide in 2022 - and if you assume further growth, it quickly becomes clear that there are not that many qualified personnel to put all the robots into operation. More automation in the use of robots is necessary so that as many companies as possible can benefit from robotics.
Generative AI will make robot programming easier
Concrete application contexts for AI particularly concern the perception of robot systems, for example the classification of objects or the segmentation of scenes. The next push will come from generative AI, particularly for task programming. Similar to how ChatGPT generates certain outputs using prompts, the activities of robots can also be automatically converted into code as processes.
A particular challenge here compared to generating pure text results is that physical data such as forces, moments or even haptic perceptions are needed for robotics. This is the only way a robot can learn the increasingly required versatility. However, this data is difficult to find online, unlike image or text data that other generative AI tools require. To get this data with little effort means to generate physical simulations of an application will become increasingly important." [1]
1. Künstliche Intelligenz hebt die Robotik auf die nächste Stufe. Frankfurter Allgemeine Zeitung (online)Frankfurter Allgemeine Zeitung GmbH. Oct 10, 2023. Von Werner Kraus
Komentarų nėra:
Rašyti komentarą