Low Cost Sensor Based Obstacle Deteciton and Description: Experiments with Mobile Robots using Grid Representation

This article describes an algorithm for detection of obstacles, through the use of low cost sonar sensors, for the navigation of mobile robots. The use of a grid representation for the cartesian space, associated to probabilistic operations on it, allowed the implementation of a fast, simple and robust algorithm for obstacle detection. This algorithm runs always that the mobile robot navigates, and allows to overcome obstacles that appear unexpectedly in the robot's path. The information used to represent the obstacles is based in the information provided by sonar sensors. A stochastic model for the sonar sensors and their activity, associated to the platform motion, are explored by the algorithm. Some experimental results obtained through the use this algorithm are described at the end of the article.