Content

• Smart Systems
• Wireless Sensor Networks
• Communication Systems
• Performance Analysis

Project Description

Wireless sensor networks (WSNs) are increasingly used for monitoring, automation and control in cyber-physical systems such as industrial facilities. There, a WSN usually reports sensory data to a central control instance — a safety-critical task that requires predictable network performance in order to timely detect potentially hazardous deviations from normal operation.

Predicting the the actual performance of a WSN is hard. It highly depends on application requirements and the derived network setting. On the one hand, it might be necessary to have separate sensors for different tasks such as temperature monitoring of an area or reporting of a machine’s log data. Furthermore, some properties are to be captured by multiple sensors simultaneously in order to identify a single sensor’s misbehavior and to prevent erroneous conclusions. Thus, such wireless networks are usually dense and impose a high communication workload on nodes, high contention on the medium and increasing transmission delays on data flows.

Obviously, the setting derived from the application requirements (sensor density, sensor distribution, sensing frequency, etc.) will make compliance with strict real-time requirements difficult. Therefore, we will approach the design of WSNs as well as their performance evaluation in the PEDS project. The goal is to learn the basics of modeling WSNs, especially the use of their shared resources, and to predict their performance. The mathematical framework we use even allows the derivation of performance guarantees, making sure that application deadlines are never violated.

Organisation