Visual Control and Tracking of Autonomous Quad-Copter Using Camera

The objective of my thesis was to design, model, control, and experimentally study low-cost autonomous drone which can fly indoors along with onboard computation and line tracking system for educational purpose. The proposed system is expected to guide the drone to follow a track on the ground smoothly without using any GPS signals. The main aim was to design and implement a nonlinear, path following system which should involve a low-cost/low-power platform, should be simple to use, safe and easy to implement with minimal need of prerequisites. The closed-loop system is designed to make the system more stable and track the path smoothly as compared to the manual drones. Adding to the advantages, the autonomous mechanism converts the under-actuated quad-copter into an over-actuated robot which provides us with full control of the state-space. Also, the research has been done to create dynamics of quadcopter and implement it in the mechanism so as to achieve the target, stability and make the quad-copter immune to any external noise. Also implemented a line detection algorithm that is capable of detecting the average position and orientation of the lines captured by a down-facing camera interfaced to the drone. The design of the altitude control algorithm using four down-ward facing ultrasonic sensors is created to provide extra stability and hovering mechanism to the quadcopter and prevent the drone from drifting and to reduce the drifting issues.