Dynamic Task Assignment of Multiple Heterogeneous Unmanned Aerial Vehicles based on Consensus with Uncertainties

This paper considers the problem of maximizing the task-assignment reward of a fleet of heterogeneous UAVs for a dynamic reconnaissance and confirmation task under uncertain scenarios and multi-UAV tasks where the coupled path optimization objectives also need to be considered. The existing consensus-based bundle algorithm is extended with an effective method for managing the multi-task and multi-agent constraints; the Bayesian estimation is then adopted to deal with uncertainties in the given scenario. The proposed method is verified by the sample run tests on a disaster area reconnaissance and confirmation task. The results verify both the practicality and advantages of the proposed method. A robust extension to consensus-based bundle algorithm that handles coupling with path planning optimization in dynamic search and rescue scenarios given tasks with multi-UAV service requirement and time-critical constraint is created.