Scenarios for Physical Robot-Environment Interaction

In this paper, we focus is on the interaction between robot and the dynamically changing environment. According to the proposed approach, a collision is determined and an appropriate scenario of behavior and response is chosen, considering the nature of interaction and location of the contact point. The neural network is used to classify the collision, and the finite automaton selects appropriate reaction algorithms. The developed approach was tested on the robot Kuka LBR IIWA 14.