Visual SLAM and Obstacle Avoidance in Real Time for Mobile Robots Navigation

Abstract

An important objective of an autonomous vehicle is to navigate through an unknown environment. A method used to achieve this objective is to generate a map. A map provides the means for the vehicle to create paths between the visited places autonomously in order to perform a task. A particular problem is to obtain such a map when there is no initial knowledge of the surroundings or not even the initial position of the robot in the environment. On other hand, avoiding static and dynamic obstacles is required, so a novel artificial potential field method is presented. The new designs that solve both problems are implemented on an FPGA. The novel designs are then tested on differential traction mobile robots with a computer vision system that travel on a controlled unknown environment. The experimental results show good performance in real time.