Adaptive UAV swarm mission planning by temporal difference learning

The prevalence of Unmanned Aerial Vehicles (UAVs) in precision agriculture has been growing rapidly. This paper tackles the UAV global mission planning problem by incorporating a greater capacity for human-machine teaming in the architecture of a flexibly autonomous, near-fully-distributed Mission Management System for UAV swarms. Subsequently, the two problems of global mission planning are solved simultaneously using an integrated solution. This consists of a geometric clustering algorithm which prioritizes the minimization of overall mission time, and an off-policy, model-free Temporal Difference Learning global agent capable of learning about an initially unknown mission environment through simulations. The latter component makes the solution adaptive to missions with different requirements.