Modified SOM based intelligent semi-autonomous navigation system

Abstract

The fundamental problem in today's world when robots are becoming a part of daily routine is getting them to move from one desired place to another without any problem and collision. In doing so a robot needs to perceive its environment to efficiently navigate through it. The navigation is possible only if the robot differentiates between different scenarios around it in real time. In this paper we present a semi-autonomous navigation strategy designed for low throughput interfaces. The task is to minimize the user input by intelligently interpreting and executing the commands. Analyzing the environment in real time, the robot selects the most probable action using collision free steering techniques which employ the Self Organizing Map in cooperation with Artificial Potential Field, in which both the user and an intelligent robot co-exist. The performance of the proposed navigation algorithm is confirmed with computer simulations and experiments using the Pioneer 3-AT mobile robot.