Machine Learning

Machine learning has revolutionized computer vision and robotics in recent years, achieving breakthroughs in both fields. In robotics, machine learning enables robots to better understand their environment and make decisions autonomously. By analyzing sensor data from devices such as laser scanners or cameras, robots can recognize obstacles and thus improve their ability to navigate autonomously, identify objects or perform more complex tasks. This is particularly useful in industry, where robots are used in manufacturing, logistics and maintenance.

Machine learning has brought about a profound change in image processing, significantly expanding the possibilities for analyzing and interpreting visual data. In particular, convolutional neural networks (CNNs) make it possible to automatically extract relevant features from raw image data. This technology can be used to recognize patterns, segment objects and even analyze emotional expressions in images. ML is typically used in medical image diagnosis, where algorithms detect diseases in X-ray or MRI images, or in logistics, where it is used for image-based tracking of goods or for object recognition and counting, label recognition, and multi-object tracking in logistics yards or warehouses.

Another focus of ours is on Large Language Models (LLMs), a subfield of generative AI. We use LLMs to integrate cognitive abilities into robotics. These models do not serve as a human-machine interface, but rather enable robots to understand human-like thought processes and make context-sensitive decisions. In this way, we are expanding the spectrum of robotics to include an intuitive, interactive component that enables versatile interaction options and a high degree of adaptability.