An increasing number of sensors in production machines promises helpful data about industrial processes, and valuable insights. Collecting data from these sensors is often not easy, though, especially in legacy environments. Wireless sensor data collection protocols promise low deployment costs and high flexibility. However, wireless communication in industrial environments is inherently challenging. In this talk, we look at a use case from the injection moulding industry, and discuss techniques to efficiently and effectively collect sensor data from injection moulding machines via a wireless multi-hop network. This includes data transformation and prioritization at the source as well as network coding for increased robustness during data transmission. We also outline a new evaluation tool which we developed to support the design of the protocols; it allows to run unmodified (binary) production code in the ns-3 discrete event network simulator.

Speaker bio:
Björn Scheuermann is Professor of Computer Engineering at Humboldt University of Berlin, Germany, and a research director at the Humboldt Institute for Internet and Society (HIIG). He holds a B.Sc. in Mathematics and Computer Science, a Diploma degree (German M.Sc. equivalent) and a PhD in Computer Science. After professorships at the universities of Düsseldorf, Würzburg and Bonn he joined Humboldt University in October 2012. The focus of his scientific work is on performance, design, and security aspects of computer networks.