The automatic motion or trajectory planning is essential for several tasks that lead to the autonomy increase of Unmanned Aerial Vehicles (UAVs). This work proposes a Dijkstra algorithm for fixed-wing UAVs trajectory planning. The navigation environments are represented by sets of visibility graphs constructed through the terrain elevations of these environments. Digital elevation models are used to represent the terrain elevations. A heuristics to verify if a trajectory is collision-free is also proposed in this work. This heuristics is a method of grid-based local search which presents linear computational time O(n p), where n p is the number of verification steps. This heuristics is compared with another method for collision verification. Results are presented in this work.

Path planning is one of the most important process on applications such as navigating autonomous vehicles, computer graphics, game development, robotics, and protein folding. It ensures that a path is planned between an initial and final position on the collision-free region of a search space if one exists. One of the most wide algorithms used for this purpose is the rapidly-exploring random tree (RRT), in which each node of a tree data structure is generated from a search space by a random sampling process, which originally follows a uniform spatial distribution. However, some authors claim that the addition of a non-uniform/informed approach into the sampling process of the RRT could accelerate the planning time of the algorithm. Actually, many works on literature propose different strategies to include non-uniform/informed behavior on RRT-based algorithms. However, the large number of studies on path planning subject impose difficulties on the identification of new solutions on a review process. The aim of this paper is to structure a review process to deal with the massive volume of works on this subject, by presenting the planning, development, and results of a systematic literature review (SLR), to investigate non-uniform/informed sampling solutions applied to RRT-based algorithms on path planning literature. A review protocol with two scientific questions was developed to guide the investigation. As a result, 1136 studies were selected in the path planning literature, of which 53 were identified as claiming to contain a solution with non-uniform/informed sampling on RRT-based algorithms. As a specific work is considered a scientific contribution only when it has not yet been explored in scientific circles, the results of the SLR can be used as a tool to search for what has not yet been proposed, helping to identify opportunities to contribute with new sampling processes of RRT-based algorithms. To the best knowledge of the authors, this paper presents the first development of an SLR of a topic related to the RRT algorithm. INDEX TERMS Non-uniform sampling, informed sampling, path planning, RRT, systematic literature review.

GERALDO MULATO DE LIMA FILHO received the B.Sc. degree in aeronautical science from the Air Force Academy (AFA), Brazil, in 2001, and the M.Sc. degree in science and space technologies from the Aeronautical Technology Institute (ITA), Brazil, in 2015, where he is currently pursuing the Ph.D. degree. Besides, he has been working as a Pilot at the Brazilian Air Force (FAB), for over 20 years. His research interests include decision support systems, MAV/UAV cooperative engagement, computational optimization techniques, and applications of artificial intelligence methods.ANDRÉ ROSSI KUROSWISKI received the B.Sc. degree in aeronautical science from the Air Force Academy (AFA), Brazil, in 2004, and the B.Sc. degree in electronic engineering and the M.Sc. degree in science and space technologies from the Aeronautical Technology Institute (ITA), Brazil, in 2017 and 2019, respectively, where he is currently pursuing the Ph.D. degree. Before engaging in research and development projects, flew for ten years in the Brazilian Air Force, carrying out various types of missions, from air defense, as a Fighter Pilot, to transport missions in support of Amazonian forest integration and protection, from 2002 to 2012. His research interests include modeling and simulation for aerospace scenarios analysis, autonomous agents, machine learning, and computational optimization.

No abstract

Nowadays, research has been developed in order to increase the autonomy of Unmanned Aerial Vehicles (UAVs). The procedure of autonomy increase consists of transferring part of the decision-making process of the human UAV operators for the vehicle itself. This work approaches the autonomous landing of UAV of the type Vertical Takeoff and Landing (VTOL). The VTOL UAV autonomous landing is a complex problem due to the existence of a significant error between the real helipad's position and the helipad's position estimated by the UAV navigation system, when this system is based on the fusion of data from a Global Positioning System (GPS) receptor and measurements from an Inertial Navigation System (INS). Thus, the objective of this work is the development of a computer vision system for tracking helipads in digital images obtained in outdoor environments, with nadir direction. Through information of the tracked helipad, the system estimates a set of flight data and generates commands to the UAV's autopilot, in order to reduce the error between the helipad's central position and the UAV's position.