Biofeedback is used in a variety of medical and technical applications to record and analyze the body's own signals. Significant achievements in sensor technology have enabled continuous further development of the underlying systems in recent years. The precise and continuous recording of biological processes is a challenging task, especially as these can vary greatly from individual to individual. In addition, interactions between the individual signals can make recording and evaluation difficult. In order to use biofeedback data in the field of human-robot interaction, key influencing factors must be recorded and models derived to enable safe and error-free cooperation between humans and robots.

Biodata for robot customization

In the context of human-machine interaction, contrary to medical and psychological applications, bodily functions are not reported back to humans, but made accessible to a machine. The original therapeutic approach of biofeedback is therefore no longer being pursued. Instead, the biological data should help to improve the interaction between humans and machines. In the collaborative assembly application, a robot should act in a way that is adapted to the employee. Special methods for measuring biofeedback are available for different biological processes. These are usually used in combination. A distinction is made between the measurement of electrical muscle activity (EMG), cardiac currents (ECG), brain activity (EEG), skin conductivity (EDA), skin temperature, blood circulation, respiration and biofeedback of internal organs.

Initial results from the work of the Bremen Institute of Structural Mechanics and Production Systems (bime) confirm the assumption that biofeedback can contribute to an improvement in human-robot interaction in the context of collaborative assembly. Based on an investigation of existing biofeedback applications from the fields of production, medicine and media effectiveness research, biosignal-based interaction models are derived. The diversity of biofeedback applications is underlined by modeling the various tasks of collaborative robots. Out of 13 developed concepts, five are further specified and elaborated. Each model can compensate for weaknesses in existing collaborative systems and enable safe and more user-friendly cooperation between humans and machines.

A prototype based on EDA and heart rate sensors was developed to implement a practical model. The developed prototype fulfills previously defined acceptance criteria such as intuitive operation, no manual input and automated recording and evaluation of body values. Communication between humans and robots is supported by visual feedback using LED bars.