Fuzzy help in mobile robot navigation

Abstract

To control the motion of a mobile robot in a partially-known environment, a hybrid method is used to fully utilize the advantages of both global and local strategies. The pre-acquired knowledge of the environment is memorized. It is useful for generating a global route and also for comparison between the real scene and the memorized one. If the sensors data in both scenes are the same, the navigation is then carried out at high speed with the assistance of a local fuzzy controller. If not, the robot is in a locally unknown environment. Then a reactive method, based completely on sensors information, such as the potential fields, is applied to navigate at low velocity. Due to the fuzzy logic, a module based on the human decision abilities is built to incorporate uncertainties on sensors data. It permits on one hand to switch between these two strategies and on the other hand to give the setpoint of robot's forward speed. The latter is modulated as a function of the output value of this module.