Computer vision for an autonomous mobile robot

Abstract

Computer vision systems are essential for practical, autonomous, mobile robots – machines that employ artificial intelligence and control their own motion within an environment. As with biological systems, computer vision systems include the vision sensors as well as the processing and interpretation of the sensor data. Therefore, a computer vision system will have one or more primary vision sensors, such as a set of cameras or a laser scanner, as well as sophisticated computer algorithms for extracting information from the raw sensor data. In Withey’s talk, methods suitable for computer vision in autonomous, mobile robots will be described and results from the application of these vision techniques are provided, specifically in a robot system that performs autonomous exploration and mapping. This application requires the machine to navigate within a target environment, avoiding obstacles, and to produce a map. Other applications, such as remote reconnaissance and remote inspections, will also be discussed.