Abstract
Physical human-robot interaction (pHRI) is a field of research progressing rapidly over the last years. One of the topics studied in this context is safety. So far, safety has been mainly addressed in relation to impacts over a single interaction point and by limiting power and energy flowing over this point. In contrast, in our work we addressed multiple, intermitted and changing contact points. For this purpose, we developed an energy-based model of the system involving both human and robot. Subsequently, this model was integrated with a controller that considers power and force information related to the interaction points and shapes energy and power flows within the system to enhance safety, while still keeping the nominal controller in the loop and responsible for fulfilling the task at hand as much as possible. Additionally, the thesis proposed a methodology for estimating human arm endpoint forces directly from muscle activations, using a motion capture system with an underlying biomechanical model of the human arm. This approach enables the estimation of interaction forces exerted by the human at contact points without the reliance on sensors such as force/torque sensors.