Sistema de Navegação Autônoma para Drones de Asa Rotativa

Abstract

The use of service robots in place of human workers has several advantages for indoor tasks, such as industrial environments and production plants, especially scenarios that pose risks to worker safety. However, the use of robots in these tasks faces significant difficulties, such as the unavailability of GPS signal and limited accessibility. This article proposes an autonomous navigation system for rotary-wing aerial vehicles capable of dealing with these difficulties. Our proposed system uses a variety of sensors, allowing the application of different localization and mapping techniques. The navigation strategy consists of traveling along a desired path, where the robot must reach a series of target positions defined by the operator. To do this, the system relies on a path planner capable of generating a path that passes through all the target points, used as a reference by the vector field navigation control. The proposal is evaluated through simulations, demonstrating the ability of the proposed system to navigate the aerial vehicle effectively in a closed environment, in this case a cave. The results also illustrate the localization errors of the different SLAM systems investigated, and their effects on navigation control feedback.