A flight time approximation model for unmanned aerial vehicles: Estimating the effects of flight dynamics and wind
Henchey, Matthew J.
MetadataShow full item record
In any vehicle routing problem it is important to accurately approximate travel distances and times as these will directly impact the overall mission cost and effectiveness. In particular, unmanned aerial vehicles (UAVs) require the effects of flight dynamics and wind on the UAV to be considered as flight time and fuel consumption rates are impacted. Most current UAV mission planning and dynamic resource management models are not taking into consideration these two important aspects of flight time and are instead using straight line distances to calculate flight times. However an area of concern is the nature in which accurate flight time approximations can be calculated, as both the mission planning and dynamic resource management models require hundreds if not thousands of flight time approximations to be pre-computed. The dynamic resource management model in particular requires these flight time approximations to be computed very quickly. In this thesis three approximation models are considered; a full nonlinear simulation of a UAV and a simplified Dubins' simulation, and the Dubins' set. We determined that the best model was the Dubins' set based approximation model programmed in Java. Using this best approach a loitering task is added into the model which determines the expected flight time into and out of the loitering task. A radius of sight task and radius of avoidance task were also developed that use the idea of limited sensor range to allow the UAV to complete a task at a specified distance away from the actual waypoint as well as to avoid a given area of threat. To account for wind effects the assumption is made that the UAV will be placed in a velocity hold and therefore will accelerate or decelerate in order to compensate for wind. This means that while the flight time will not be affected the fuel consumption rate will increase or decrease depending on the direction and speed of the wind. Data was collected from a full nonlinear simulation of a UAV and a simplified fuel consumption rate model was developed and added into the flight time approximation model. This is necessary as the mission planning and dynamic resource management models make assignments based on the remaining fuel after a task has been completed. Finally the performance of the completed model is studied to show that it is able to meet the specified standards. Possible extensions and future work are also discussed.