Model predictive control strategies for parcel relay manoeuvres using drones

Abstract

The increasing supply and demand by customers and companies is an important factor present in today’s economy. With new business tactics like online shopping, the number of products being requested by clients is rapidly increasing, and to have a profitable business companies must be able to reach the client in a quick, effective way. Current delivery services distribute large quantities of items in urban environments, however, for individual deliveries, they are not the most efficient, since they result in high costs. Autonomous electric vehicles like quadcopters offer a new approach to the delivery paradigm, being more efficient and accessible, less polluting, among other advantages, however their low battery life presents a big disadvantage, limiting the range these aircrafts can reach to deliver packages. The present paper addresses this problem by studying advanced control methods, particularly Model Predictive Control (MPC) based techniques due to their effectiveness handling agile dynamics and performing aggressive manoeuvres, accounting to physical and mechanical constraints present in the process. The main goal is having two aircrafts perform a relay manoeuvre, exchanging a payload between them and effectively extending the delivery range. Firstly, the vehicle’s model is established and used to formulate the optimal control problem. The resulting controller will be subjected to validation using complex trajectories and compared to an implementation of another similar controller. After validation, the controller will be adapted to support two vehicles simultaneously and subjected to a parcel exchange scenario. Conclusions about the feasibility of the controller will be drawn.