Experimental validation of quadrotor simulation tool for flight within building wakes

Experimental test flights are conducted to validate a novel simulation tool capable of predicting the effects of building generated turbulent winds on the flight performance of autonomous quadrotors. Wind conditions are modeled using time accurate Large Eddy Simulations where five representative points within the building wake are used to predict an average position hold deviation. Three control methods for autonomous position control tuned under no wind conditions are evaluated (Proportional Derivative (PD), Integral Backstepping, and Fuzzy Logic). Flight tests using a small (≈2 kg) quadrotor UAV within a building wake show that the simulation tool is able to predict average position deviation to within a single body length (0.55 m) in the horizontal plane for observed freestream wind speeds between 1 and 2 m/s (with predicted gusts of up to ±6 m/s). The simulation tool is also shown to be able to predict relative differences between control methods both overall and specifically along each co-ordinate axis. A PD controller is predicted to hold position more accurately with a building wake in the direction perpendicular to the freestream wind by ≈± half a body length while a Fuzzy controller achieves the same result in the direction parallel to the freestream wind. These results are validated by flight test data where both the most effective control strategy, and the magnitude of the position control difference, are shown to match the simulation results.