Energy-aware coverage path planning for unmanned aerial vehicles
The Coverage Path Planning (CPP) problem is a motion planning subtopic in robotics, where it is necessary to build a path for a robot to explore every location in a given scenario. Unmanned Aerial Vehicles (UAV) have been employed in several application domains related to the CPP problem. Despite th...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2019 |
| País: | Brasil |
| Institución: | Universidade Federal de Pelotas (UFPEL) |
| Repositorio: | Repositório Institucional da UFPel - Guaiaca |
| Idioma: | portugués |
| OAI Identifier: | oai:guaiaca.ufpel.edu.br:prefix/6405 |
| Acceso en línea: | http://guaiaca.ufpel.edu.br/handle/prefix/6405 |
| Access Level: | acceso abierto |
| Palabra clave: | CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO Computação Coverage path planning Energy-aware UAV Flight pattern Complete algorithm Pheromone-based heuristic Planejamento de caminhos de cobertura Preocupação energética VANT Padrão de voo Algoritmo completo Heurística baseada em feromônio |
| Sumario: | The Coverage Path Planning (CPP) problem is a motion planning subtopic in robotics, where it is necessary to build a path for a robot to explore every location in a given scenario. Unmanned Aerial Vehicles (UAV) have been employed in several application domains related to the CPP problem. Despite the technological progress related to control systems and energy monitoring, one of the most significant limitations of UAVs is endurance, especially in multirotors. These vehicles have a limited payload, which consequently also restricts their batteries. In this way, minimizing the energy consumption of these vehicles is pivotal to prolong and guarantee the coverage mission. The energy consumption is usually associated with the number of turning maneuvers in a given trajectory. Most studies in literature seek to minimize this issue to save energy and, consequently, enhance endurance. However, important factors are usually ignored, such as the acceleration and deceleration phases, optimal speed to travel a straight distance, the entrance speed and turning angle when performing a maneuver. Thus, this work aims to propose energy-aware coverage path planning solutions based on flight patterns, complete algorithms, and pheromone-based methods for regular and irregular-shaped areas containing full and partial information considering the impact of different aspects in the UAV energy consumption, such as traveled distance, mission execution time, turning maneuvers, optimal speed, and external conditions. First, we present an energy-aware spiral coverage path planning algorithm called E-Spiral. The flight pattern performs missions over regular-shaped areas consisting of concave and convex polygons. The proposed approach considers specific requirement applications, such as overlapping and image resolution, to guarantee a complete area mapping in photogrammetric sensing applications. Furthermore, the algorithm explores an improved energy model to adopt optimal speeds in straight segments of the path to minimize the total energy consumption. Next, we present an energy-aware grid-based coverage path planning algorithm called EG-CPP. The complete algorithm generates coverage paths for mapping missions over irregular-shaped areas containing no-fly zones. Our solution improves an existing grid-based method by replacing its original cost function based on the sum of angles to an energy-aware cost function. As a further contribution, two pruning techniques improve the performance of the algorithm. This improvement speeds up the computational time of the algorithm up to 99%, allowing to explore all the different starting positions in the workspace and saving even more energy. Then, we present an energy-aware pheromone-based solution for patrolling missions called NC-Drone. The approach extends an existing real-time search method aiming at minimizing the number of turning maneuvers while keeping the unpredictable behavior during coverage. We propose two types of NC-Drone, a centralized algorithm and a decentralized one with a few variations. These variations explore a matrix-representation to store the visited cells in the vehicle’s memory and adopt synchronization schemes to share the information between the UAVs. We also propose cooperative strategies to improve the algorithm and further explore the problem considering relevant aspects, such as time, uncertainty, and communication. We combine all NC-Drone strategies into a single solution for the patrolling problem. Therefore, three novel approaches are proposed able to successfully addressing several problems related to different coverage path planning scenarios, advancing the state-of-the-art in this area. |
|---|