Tracking Stream Quality Issues in Combined Physical and Radar Sensors for IoT-based Data-driven Actuation

In this paper, a stream quality tracking for measurements from combined radar and physical sensors is developed. The authors proposed the use of RDF stream processing system and semantic rules to provide semantic reasoning for tracking erroneous data points from real time sensor readings. We demonstrated the effectiveness and the efficiency of the approach using the dataset produced from smart home project. Experiments were conducted with simulated arbitrary sensor values for inconsistent and missing sensor readings at various data points. The results was able to show the feasibility of the approach and that quality requirement of sensor streams can be verified at lower granularity level (data layer) of smart actuators.