Prediktivní řízení autonomní helikoptéry s časově proměnou hmotností

Abstract

The ultimate purpose of this thesis is to bring closer to reality, the use of fully autonomous unmanned aerial vehicle to extinguish wildfires. In order to achieve that goal, the approach taken consists in designing a solution to an optimal control problem for which a UAV can deploy a payload (e.g. re retardant or water) above a wildre. This solution also avoids structural damage to the UAV (due to heat exposure), and minimize the re retardant dissipation. This is achieved by minimizing both and the releasing distance from the re and by maximizing the speed upon releasing the payload. In addition, the optimizer takes into account environmental parameters such as wind and terrain gradient. The novelty in this method is to deliberately cause the drone to fly outside of its safe flight envelope. This means, by tilting the UAV to high pitch angles, allowing it to engage in a sort of controlled free fall, while the thrust is pointed almost horizontally, and thus achieving higher horizontal speeds than it would normally be able to. By doing so, the high heat exposure time can be minimized and the payload can be dropped closer to the fire epicenter. The optimization process takes into account the expected change in mass (due to the payload release above the re), and allows it to engage in a risky maneuver, assuming that after dropping the payload, the vehicle will be lighter and thus able to recover without impacting on the terrain. The output of the optimizer consists of a full state trajectory for the whole planned maneuver. These outputs were tested with both simulators and real platforms.

The ultimate purpose of this thesis is to bring closer to reality, the use of fully autonomous unmanned aerial vehicle to extinguish wildfires. In order to achieve that goal, the approach taken consists in designing a solution to an optimal control problem for which a UAV can deploy a payload (e.g. re retardant or water) above a wildre. This solution also avoids structural damage to the UAV (due to heat exposure), and minimize the re retardant dissipation. This is achieved by minimizing both and the releasing distance from the re and by maximizing the speed upon releasing the payload. In addition, the optimizer takes into account environmental parameters such as wind and terrain gradient. The novelty in this method is to deliberately cause the drone to fly outside of its safe flight envelope. This means, by tilting the UAV to high pitch angles, allowing it to engage in a sort of controlled free fall, while the thrust is pointed almost horizontally, and thus achieving higher horizontal speeds than it would normally be able to. By doing so, the high heat exposure time can be minimized and the payload can be dropped closer to the fire epicenter. The optimization process takes into account the expected change in mass (due to the payload release above the re), and allows it to engage in a risky maneuver, assuming that after dropping the payload, the vehicle will be lighter and thus able to recover without impacting on the terrain. The output of the optimizer consists of a full state trajectory for the whole planned maneuver. These outputs were tested with both simulators and real platforms.