Without programming: Robots learn gripping movements from humans TU Braunschweig spin-off presents prototypes
A spin-off company of Technische Universität Braunschweig is working on robots that learn gripping movements directly from humans. The “HandEffector” has a robot hand modelled on a human. It can detect gripping forces via a sense of force and imitate them using artificial intelligence. In this way, the human can train the robot by demonstrating gripping movements only once. This fundamentally changes robot programming, making it particularly intuitive. The founding team is now presenting a prototype that is being tested and expanded to include functionality.
The “HandEffector” system was developed to reduce the implementation costs for new processes: the intuitive learning method is intended to speed up the programming process. Initial experiments show that more than 95 percent of the learning time can be saved. Through integrated force measurement, AI-based object recognition and grip adaptation, the system offers automated handling of different, pressure-sensitive objects such as food, as well as safe interaction with humans. This enables the automation of processes that previously could not be automated with existing robotic solutions. Finally, integration into existing production lines is easy as object recognition and AI are already included in the system.
The HandEffector system is a cobot, i.e. a collaborative robot. It is designed for use in cooperation with humans. Safety cells as in classical large industrial robots, e.g. in car production, are not required. Each joint of the robot monitors independently acting forces and thus complies with safety limits for collaborative work with humans.
Possible Areas of Application
Initial applications are in the food industry: The system can be used in the sorting and packaging of sensitive products. As a future application, it is planned to use the system as a service robot to support building cleaning and maintenance.
The concept supports employees in monotonous, repetitive and physically harmful work steps. But the lack of personnel in certain production and service sectors also motivates the use of robot technology. In the spirit of the EU’s “empowerment” approach, workers are also to be enabled to perform tasks that they were previously unable to do.
The founding team consists of two former employees of the Institute of Robotics and Process Control at TU Braunschweig (the brothers Dr.-Ing. Sönke Michalik and Dr.-Ing. Sören Michalik) and a former employee of the Chair of Business Administration and Corporate Management at TU Clausthal (Dr. Lars Heim).
The founders, Sönke and Sören Michalik, worked at TU Braunschweig’s IRP Institute until the spin-off was founded. They designed the drive elements used in the project and developed them to product maturity.
Sönke Michalik was responsible for the management and implementation of technical developments in the field of modular robotics in European joint projects. As part of his doctorate and in other research projects, he researched the hardware acceleration of image processing using artificial intelligence, FPGA-based stereo camera systems and latency-free data transmission.
Sören Michalik has researched and developed hardware accelerators for satellite communication and data processing in a project of the European Space Agency (ESA), and other EU projects. In his doctoral thesis, he worked on hardware-accelerated robot localisation and navigation and developed a subpixel-accurate method for determining image features.
Lars Heim holds master’s degrees in marketing and business psychology and completed his doctorate at TU Clausthal in 2021 at the Chair of Business Administration and Corporate Management with a focus on business models, particularly in relation to blockchain technology. Through various positions in business, he also has practical experience in marketing, project management and sales, among other things. Through the founding of a start-up for the product development, production and distribution of a kitchen appliance in 2016, he continued to gain initial experience in the area of company formation.