Robot workspace monitoring with redundant structured light cameras - A preliminary investigation

In this paper we propose the use of a redundant array of structured light scanning cameras to monitor a collaborative robot workspace. We present the model and suggest a minimum number of such cameras required to monitor a particular area. We then propose a concept for segmenting the workspace into different sub volumes to allow for different categories of obstacles. We then propose that a voting scheme will allow us to process multiple camera inputs in real-time in a safe fashion. We perform initial experiments and draw appropriate conclusions before defining further work.