Spectrum and AI-based Analysis for a Flight Environment and Avoiding Virtual Obstacles Using Potential Field Method for Path Control

Computer vision-based navigation systems basically rely on the external environment for positioning. This research studies the selection of the computer vision navigation algorithm that suits better for a specified flight environment. Two methods are used to analyze the flight environment images, and a com- parative study between them is presented. While the first method depends on spectrum analysis using Fourier transform offline, the second method uses artifi- cial intelligence based on Convolutional Neural Network (CNN) to analyze such images. Avoiding the bad matching areas and treating them as virtual obstacles is realized using the potential fields method. The two methods are implemented and tested on a path in a simulation environment consisting of Robot Operating system (ROS), Gazebo simulator, and IRIS drone model. Results show that both methods give good indicators to the more efficient navigation algorithm, but CNN offers more detailed description of the environment and additional options to avoid failures.