Cooperative distributed trajectory optimization for a heterogeneous UAV formation

In this paper, we formulate a mathematical cooperative distributed trajectory optimization approach using a receding horizon strategy, as a Mixed Integer Linear Programming (MILP) problem for a heterogeneous formation of aerial vehicles with independent dynamics but coupled objectives. We propose a realistic scenario, in which a ground network of static sensors has to send data to a Ground Network Control Server (GNCS), using a Low Power Wide Area Network (LPWAN) technology, through a formation of fixed-wing and multirotor Unmanned Aerial Vehicles (UAVs) acting as relay gateways. The UAVs and the GNCS are linked using a wireless cellular network 4G LTE/5G connection. Each member receives its own waypoints describing the trajectory from the GNCS and seeks for an optimal fuel (electrical power) solution of a derived objective function, in the context of low transmit-receive data packets failures. Object and inter-member collision equations are formulated as MILP constraints for the objective function. Channel communication failures, because of fading, are characterized using a Rice model and are added to the overall mathematical problem. Finally, the cost function to minimize is obtained, upon which a numerical simulation is performed through the use of a commercially available MATLAB MILP solver such as INTLINPROG function.