Implementação de um Planejador de Caminhos Embarcado num Robô Móvel Navegando em Terrenos Acidentados

Abstract

The use of mobile robots in performing services is becoming increasingly common. One example is Vale’s EspeleoRobô, capable of moving over rough terrain to inspect and map confined environments, such as piping systems, dam drain galleries, mills, silos, and natural caves. Due to space limitations, the presence of obstacles and terrain irregularities, teleoperation via radio waves or umbilical cable is not always a viable option. A possible solution corresponds to the development of strategies that allow the autonomous operation of such robotic devices. Among these strategies, path planning is essential to ensure that the robot travels efficiently to the desired locations while avoiding collisions with obstacles. In this context, this article presents the implementation and experimental validation of the Dijkstra path planning algorithm, integrated with auxiliary techniques of terrain modeling by voxels, and mobility metrics adopted to define the paths’ costs. Programs implemented with ROS were embedded into the EspeleoRobô, enabling to accomplish experiments in the laboratory and in the field. The obtained results demonstrate the proper functioning of the implemented algorithms, enabling its integration with the robot autonomous navigation system.