Consideration of communication in human-machine interaction for cooperative trajectory planning

Julian Schneider ^* Balint Varga Sören Hohmann

• Institute of Control Systems, Karlsruhe Institute of Technology, Karlsruhe, Baden-Württemberg, Germany

Interactive human-machine systems aim to significantly enhance performance and to reduce in the workload for humans by leveraging the combined strengths of human and automated systems.In the state of the art, human-machine cooperations are modelled in various interaction layers, e.g. decision layer, trajectory layer and action layer. The literature usually focuses on the action layer, assuming that there is no need for a consens on the decision or trajectoy layers. Only few works deal with the interaction on the trajectory layer above this. Non of the previous work has addressed the systematic structure of communication for interaction between humans and machines besides the action layer. Therefore, this paper proposes a graph representation based on multi-agent system theory for human-machine cooperation. For this purpose, a layer model for human-machine cooperation from the literature is converted into a graph representation. Using our novel graph representation, the existence of the communication loops can be proven, which are necessary for an emancipated cooperation. In contrast, a leader-follower structure does not posess a closed loop in this graph representation. The choice of the communication loop for emancipated cooperation is ambiguous and can take place via various closed loops for humanmachine interaction at higher layers which open new possibilities for the design of emancipated cooperative control systems. In a simulation, it is shown that emancipated cooperation is possible via three variants of the communication loops and that a consensus on a common trajectory is found in each case. The results indicate that taking into account the cooperative strategies at the trajectory layer can enhance the performance and effectiveness of human-machine systems.