SLAM com associação de dados conhecidos: uma aplicação usando EKF e mapa com grade de ocupação

Abstract

Simultaneous location and mapping (SLAM) has been a topic of much research over the past two decades in the computer vision and robotics communities, and has recently received the attention of high-tech companies. SLAM techniques create a map of an unknown environment and locate the robot on that map, with a strong focus on real-time operation. Among the different sensor modalities, cameras are inexpensive and provide rich information of the environment that allows for robust and accurate local recognition. Therefore, visual SLAM solutions, in which the main sensor is a camera, are of great interest. This work presents an implementation of simultaneous location and mapping (SLAM) based on the extended Kalman filter (EKF). For the implementation, the programming language python, the set of libraries and tools of the robot operating system (ROS) and the computer vision library OpenCV were used. The work was developed in the open source simulator Gazebo, where we created an environment for the robot, simulated the robot in that environment, and then built a map of that environment. The result was compared with the original map and found to be relatively accurate. The resulting map shows some inaccuracies, but the overall shape is acceptable.