Ph.D Thesis

Ph.D StudentPotyagailo Svetlana
SubjectPlanning and Operational Algorithms for Autonomous
DepartmentDepartment of Aerospace Engineering
Supervisors PROFESSOR EMERITUS Omri Rand
DR. Yaron Kanza
Full Thesis textFull thesis text - English Version


This research is focused on targeted flight of an unmanned helicopter in a GPS-denied environment and unknown map. This effort includes the development of the methodology and the required algorithms for such missions where the vehicle faces different challenges, such as the absence of accurate information regarding its position, lack of a priory known map of the environment, and unavailability of additional aid from external systems.

We have developed a modular system for solving this task for a conventional micro-helicopter. This system tackles different aspects of an autonomous flight including estimation of the vehicle's position and orientation, creation and updating of a virtual map of the environment, planning a safe path towards a target, and computing the commands required to fly along the planned path. The system consists of three main modules that include novel algorithms and techniques to solve the above mentioned tasks. Using the principle of modularity and adopting appropriate scaling laws, the system can be adapted to a wide range of configurations, including mini-scale and full-scale, indoor and outdoor rotary-wing autonomous vehicles.

As part of the proposed system, we have developed a method for localization of a vehicle within an unknown environment simultaneously with the creation of the map of that environment. This method is model-free and merely employs a laser range finder, for sensing and detecting obstacles in the environment. The proposed path planning algorithms are aimed at planning a feasible, safe and collision-free path, from a given initial position and orientation to a specific goal location, while taking into account the maneuvering limitations of the vehicle. Additional components of the system include a method for computing smooth flight trajectories and required control commands. To simulate the entire system, we have developed a detailed and high fidelity simulative helicopter model for testing the system in different environments. Our simulation results illustrate the effectiveness of the proposed approaches and the variety of sub-problems that can be solved by additional combinations of our system's modules.

The primary use of our system may be in the field of search and rescue missions, where flying vehicles may be exploited to explore dangerous areas or to localize and examine hazardous or remote objects. However, it is important to note that the proposed system may be used in a wide range of missions as demonstrated throughout this work.