Toronto Metropolitan University
Adhikari_Min_Prasad.pdf (11.5 MB)

Real time autonomous collision avoidance for unmanned aerial vehicles

Download (11.5 MB)
posted on 2021-05-24, 19:02 authored by Min Prasad Adhikari
GeoSurv II is a jointly funded project of Sander Geophysics Limited (SGL) and NSERC to develop a fixed-wing Unmanned Aerial Vehicle (UAV), capable of autonomously performing high resolution geophysical surveys at low flight altitudes over poorly known terrain. This thesis is in support of achieving this objective. In order to achieve such a level of autonomy, the UAV must be capable of avoiding stationary, pop-up and moving obstacles while flying at low altitude. Such obstacles may include power lines, communication towers, trees, unknown flight vehicles encountered while at flight or uneven terrains which creates the situation of the pop-up obstacles. In addition to that the UAV must be able to fly as close as possible to the reference trajectory for a given geophysical survey. The development and testing of a method capable of performing such an autonomous mission is the objective of this thesis. In this thesis, a method is developed based on a spectral method known as Legendre Pseudospectral Optimal Control, because of its capability to directly incorporate all of the mission objectives, while respecting the UAV constraints (which other methods in the literature are not capable of). The method accounts the aircraft and obstacle constraints there by capable of avoiding obstacles with feasible maneuvers for the aircraft. The objective to remain as close to the reference trajectory is fulfilled by setting the area between the flight trajectory and reference trajectory as the cost of optimization of the optimal control problem. Five different scenarios presented in this thesis show the developed method's capability to avoid the stationary, pop-up and the moving obstacles successfully while remaining close to the reference trajectory.





  • Master of Applied Science


  • Aerospace Engineering

Granting Institution

Ryerson University

LAC Thesis Type

  • Thesis