Protótipo de veículo autônomo terrestre para mapeamento de ambientes industriais

Abstract

The use of mobile robots is increasingly present in the development of society, from simple tasks like domestic services to more complex ones like rescue missions. Cur rently, the main challenge of a mobile robot occurs when the robot is inserted in an unknown environment, causing its route to be recalculated several times, with the de tection of information from the environment, until reaching its final goal. Therefore, this work proposes the development of a prototype of an autonomous terrestrial vehicle capable of moving around industrial environments and mapping it. The CAD model of the vehicle was first designed in the Solidworks software and to have a low produc tion cost, its parts were produced using additive manufacturing through 3D printing. The electrical part of the vehicle has a Raspberry Pi for processing data from several sensors capable of detecting obstacles and mapping the area, in addition to actuators that allow its locomotion in the environment. To facilitate the connections between the various devices present, a printed circuit board (PCB) was developed that works as a shield, connecting the microcontroller used to the devices. For programming the algorithms, the Robot Operating System (ROS) was chosen to facilitate the connec tion between the various devices and for containing packages for implementing the mapping algorithms, such as Hector SLAM. The robot was assembled and tested in 3 environments with industrial characteristics and presented an average of 93.67% of accuracy in relation to the floor plan of the place, taking approximately 12.06min/m 2 to perform the mapping in each one of the environments.