Desenvolvimento de um algoritmo de planejamento de trajetória em ambientes desconhecidos e não estruturados para UAVs

Abstract

To make a UAV (Unmanned Aerial Vehicle) autonomous, it must perform actions without human interference in its missions. Regardless of the area of operation, several autonomous UAV missions require path planning, being able to operate in unknown, 3D, unstructured environments, and with dynamic obstacles depending on the assigned mission. Thus, it is essential that the algorithm can avoid obstacles since the environment will not always be known. As well as take into account the movement restrictions to achieve smooth curves in the planning. Several path planning algorithms can be adopted among classic, metaheuristic, or machine learning techniques. One observable aspect among the existing and most used techniques is, ”which would be the best technique to work in these environments”. The algorithms that obtain the best results are tested in 3D in the UAV simulator. A deeper analysis is performed, considering completeness, distance, time, CPU usage, memory usage, collision prevention, and robustness. The final test to validate the trajectory planning algorithms is done in a real environment.