Model driven developed machine vision system for service robotics

In this paper the modelling of the machine vision architecture ROVIS for service robotics is presented. The focus of the paper is not on the image processing algorithms used for robot environment understanding, but on the Model Driven Development (MDD) approach used for the design and implementation of the vision system. Starting from the basic requirements of ROVIS, the development process goes through three typical design phases: requirements analysis, system functional analysis and architectural design. The flow of information from the user of the rehabilitation robot to the vision system and back is modeled within the so-called shared-control framework. The proposed machine vision architecture is used in implementing the visual perceptual capabilities of the rehabilitation robot FRIEND.