Protótipo de robô móvel autônomo orientado por GPS

Abstract

One of the most important aspects about the implementation of autonomous robots/vehicles and ADAS is localization. The most popular geolocalization system is GPS, which allows position estimation of the receiver at any place in the world, but has accuracy problems on the other hand. For a more accurate representation, a relevant technique is sensor fusion, that uses multiple sensor data to generate a better description of the system if compared to the description generated by each individual sensor. The present work describes the implementation of an autonomous mobile robot that is capable of reaching a coordinate remotely given by the user, using an IMU that was implemented from three-axis MEMS sensors (an accelerometer, a gyroscope and a magnetometer) for assisting the position resolution from the GPS receiver data. The implementation was done in a embedded system, which uses a development kit based on an ARM Cortex-M4F family microcontroller and RTOS for supporting the temporization and shared resources management. The results were checked from simulated and measured results from the real and “virtual” sensors (the last one are the sensor fusion generated data), as well as the system operation, assembled in the robot chassis.