Gesture spotting for controlling a mobile assistance system for service and maintenance

Abstract

As information technology interweaves with our daily environment, new modes of interaction will be required. In this paper, we suggest a gesture-based approach and present a prototypical case study for a gesture controlled mobile assistant system for an industrial application domain. We discuss a number of design requirements, that are common to gesture-based interaction techniques, describe the resulting system architecture, as well as the training and recognition process, and an experimental evaluation. We will show that even without too much effort such a system can be realized with standard hardware and algorithms, and still achieve very robust recognition results.