An Autonomous Navigational Mobile Robot

Abstract

There has been some success with autonomous robot navigation in indoor environment over the past few years, but all of the success has come at the price of expensive robot hardware. Even an entry level navigational robot costs well into the thousands of dollars. This paper explores the feasibility of autonomous navigation with a low cost robot. To keep construction costs down, the mobile robot is constructed with inexpensive sensors. A simulator, which is an absolute necessity, is used to develop the AI. A topological map making and navigation approach is used which compensates for the limited sensor arrangement. The robot is tested in an indoor office environment measuring 24 by 16 meters. In the simulator, the robot perfectly maps out and navigates in the area. In the real environment problems occur with the primary sensors. Despite these problems, the robot is still is able to generate a map. Navigating once the map is created, the easier of the two tasks, is reliable as long as the more accurate map that is generated in the simulator is used. This paper shows that despite the limited amount of perception that the sensors provide to the robot it is still possible to perform navigation, and map making.- v-Acknowledgements