Obstacle Detecction and Avoidance by a Mobile Robot Using Probabilistic Models

This article presents the implementation of an strategy for detecting, mapping and avoiding obstacles using the mobile robotic platform DaNI 2.0. This mobile robot includes an ultrasonic sensor, which is modeled based on a Gaussian function with the aim of reaching higher precision of the distance position measured for each obstacle in the scene. An avoidance obstacle algorithm, based on reactive path planning, allows the mobile robot navigates coherently around the environment. The path planning method analyses the change in potential field around the environment (the goal provides attraction forces and obstacles repulsion forces) and it is implemented on the robot DaNI using the VxWorks system. Furthermore, the implementation of an occupation grid based on the log-odds algorithm for mapping the environment is described. The robot navigation in real indoor environments with obstacles are analyzed and discussed at the end of this document.