A Hybrid Approach to Interaction Detection in Robotics

An approach to hybrid model/non-model-based interaction detection is presented in this paper. The non-model-based aspect of the hybrid approach enables indirect inclusion of unmodelled contact task dynamics into the model-based algorithm which is used to generate expected signal values for the nominal operation of the robot. Based on experimental results, this paper discusses the applicability of the proposed approach in different types of contact tasks, as well as its possibilities in scope of Industry 4.0.