Unmanned Aircraft Systems (UAS), commonly known as drones, are becoming increasingly popular in civilian applications. However, the infrastructure to ensure the safe operations are still largely missing. In this dissertation, I addressed two problems related to a flight mission. First, I propose a flight planning system specifically for single multirotor UAS operations. The system generates a 4D trajectory at the pre-flight stage before an actual flight taking place. Safety is achieved by ensuring enough energy to finish a trip under environmental influences such as wind, as well as avoiding static obstacles. Second, I proposed a hybrid controller capable of performing trajectory following as well as single-helicopter collision avoidance. The controller enables a multi-rotor to follow the planned trajectory.
In the first part of the dissertation, I first divide the flight planning system into two components: energy demand and energy supply. The demand side is determined by the performance of the UAS, the environment, and the flight trajectory, while the energy supply depends on the power source. Then, I focused on two specific problems on the energy demand side. First, I quantify the energy performance of a UAS by developing a theoretical power consumption model for multi-rotor Unmanned Aircraft Systems (UAS). The model is derived from the helicopter literature. The model parameters are identified and validated experimentally by flying an IRIS+ quadrotor UAS. Second, I solved the trajectory generation problem by formulating it as an optimal control problem. The cost to be optimized is the energy consumption or flight duration at the cruise phase of a flight mission, taking into account 3-dimensional wind field and terrain variation. The problem is solved numerically by the Ordered Upwind Method (OUM). By combining the optimized path with timestamps, we can generate 4D trajectories typically used in aviation flight planning, and thus maximize the software compatibility between manned and unmanned aviation. Simulation shows that the generated trajectories are able to avoid obstacles such as high-rise buildings and give energy consumption estimates along the entire trajectory.
In the second part of the dissertation, I addressed the trajectory following problem with collision avoidance capability given a planned 4D trajectory. I analyzed the tracking performance of an IRIS+ quadrotor for a challenging trajectory. In the scenario, wind played a big role in the cross-track errors. The collision avoidance capability, or a safety controller, is integrated to the trajectory following controller by introducing a hybrid automaton, deciding when to start the avoidance action and what control actions to take. The hybrid controller is examined in a simple straight-line-following scenario to avoid a manned helicopter.