Trajectory Optimization for Cellular-Connected UAVs with Disconnectivity Constraint

The popularity of unmanned aerial vehicles (UAVs) has been increasing recently thanks to their enhanced functionalities and decreasing manufacturing costs. Many military and civilian applications involve UAVs for usecases/missions ranging from surveillance to search and rescue. However, despite their advantages, UAVs typically require near ubiquitous network connectivity for a successful fulfillment of their missions, which is a challenging task to achieve. In this paper, we study the trajectory optimization for cellular- connected UAVs with a disconnection duration constraint. That is, the UAV with the mission of flying from a start location to a final location needs to find a path during which it does not lose its cellular connection via one of the ground base stations (GBS) in the area more than a given time constraint. As the problem is difficult to be solved optimally, we propose a dynamic programming based approximate solution within polynomial time. In simulations, we show that the proposed approach can achieve close-to-optimal results with remarkably low computation costs and its accuracy could be increased by adding more granularity to the approximation.