Towards Safe Drone Transportation in Toronto

This thesis focuses on the development of an application capable of predicting wind velocities at any location in Toronto that can be used to improve drone flight safety. First, computational fluid dynamics (CFD) was used to study the effect of downtown building geometry on the wind field. Wind speeds and bearings from CFD were extracted, compared to local weather station measurements, and were within 12% and 8% root mean square error (RMSE), respectively. Next, a simplified CFD model was developed to efficientlymodelToronto’ssuburbanzones. This model utilized a roughness parameterization to implicitly model small buildings. When compared to weather station data,the wind speeds and bearings from this model were within 8% and 2% RMSE, respectively. In combination, these chapters outline a methodology for accurately obtaining real time, historical, forecast, and statistical wind velocities in the drone flight environment which can be used to improve flight safety and efficiency.