Comparison of Event-based Remote State Estimation Techniques for UAV Formation Control

Formation control of unguided aerial vehicles (UAVs) is a challenging task and requires the exchange of large amounts of information between the agents, typically over wireless networks. To leverage the burden on the communications network, the use of event-based instead of periodic transmissions is proposed. To exploit all benefits, the multi-agent system needs to be equipped with suitable event-based estimators relying on dynamic models of the UAVs. In this paper, we investigate how detailed the modeling approach of the quadrotor dynamics (e. g., linear vs. nonlinear) needs to be in order to perform reliable formation control. We evaluate the performance of different models in combination with appropriate event-based estimation techniques regarding the system's formation error and the distribution of triggered events with the help of numerical simulations.