Geração de trajetória em espiral e navegação com desvio de obstáculos para veículos aéreos não-tripulados

Abstract

Electric power companies usually distribute electricty using transmission lines over long distances from the origin of production to large consumer centers. To ensure continuity with the minimum interruption in power supply, there is a need for periodic inspection and maintenance on the lines and their support towers. Similarly, telecommunication companies need to perform this procedure, even if it does not involve long distances from conductive cables. These companies show interest in automated and remotely controlled inspection systems as they offer a lower cost and lower risk, while maintaining or improving quality inspection. The state of the art shows a range of solutions, ranging from robots to unmanned aerial vehicles (UAVs), popularly and commercially known as “drones”. The use of UAVs in monitoring in general, and especially for transmission lines, is a multidisciplinary process that involves aspects of aerodynamics, embedded system hardware, control system, mission generation and execution algorithms, image processing, among others. Each aspect can have several solutions. This work proposes a solution for the generation and execution of missions for the inspection of electricity transmission lines or telecommunication towers, in which the UAV follows a spiral path around the tower and has the capacity of obstacle deviation, allowing a complete view of the asset inspected. The solution consists of four main algorithms: (i) trajectory generation that creates a series of waypoints that draw a spiral shape, through parameters informed by the operator, as the distance between the center of the tower and the starting point of the navigation where the UAV is positioned; (ii) an obstacle detection algorithm determines the position of the obstacle in space by obtaining data from a set of ultrasonic distance sensors positioned around the UAV; (iii) a navigation time trajectory (re)planner allows the UAV to avoid the obstacles whose algorithm was based on a circular path formed by interpolated points and (iv) a mission manager that has the objective of scheduling the waypoints to be navigated by UAV, from both the main spiral route and the waypoints generated by the obstacle deviation algorithm. Three scenarios, using simulation software, were created to validate the proposed solution, The first one does not present any obstacle and is used as a control sample. The second scenario is composed by long obstacles arranged longitudinally and transversely to the spiral trajectory, in addition to a set of zig-zag obstacles. The third scenario is simpler and provides an obstacle with a greater extension than the UAV. The performance data of the experiments in each scenario were compared with the control scenario, including the comparison of the calculated and executed routes. As a result, it was possible to observe that performed route distance increases in comparison with the calculated one; such an increase is proportional to the amount of obstacle deviation executions. The results also shown that the solution is feasible and fully functional.