Learning Vision Based Navigation with a Smartphone Mobile Robot

Abstract

In this paper we describe a small, low cost autonomous mobile vehicle for learning and research in autonomous vision guided navigation. The vehicle consists of a mobile base that holds a smartphone. The smartphone provides the camera and the computing engine while the mobile base provides the actuators and additional sensors for autonomous interaction with the environment. The smartphone sends actuation commands to the mobile base and receives sensor data from the base over a USB connection. By taking full advantage of the high performance to cost ratio of mass consumer product technology the autonomous mobile robot built in this way lowers the entry barrier for hands-on learning of autonomous vision guided navigation. We illustrate with an example how the robot together with the Android software development environment provides a learning tool for building the complex software needed for today's and tomorrows highly automated machines.