My current research interests lie in the following areas.
Energy Data Analytics
Smart meter installations have opened up unprecedented opportunities to collect energy consumption data. Likewise, plug-level meters find their ways into more and more homes, often as part of home automation solutions. As part of our research activities, we analyze the data collected from these devices and realize novel use cases, e.g., detecting unexpected consumption behavior or predicting future energy demand.
Creating Smart Spaces
Based on the collection and consolidation of contextual data from multimodal sensor devices, correlations between the environmental conditions and the users’ situations can be established. We cater for the integration of sensing devices into smart spaces in order to enable the synergistic processing of their data. Through learning previously encountered behavior, we also investigate how to adapt the physical environment to meet the user’s preferences.
Heterogeneous Sensor Networks
Wireless Sensor Networks are commonly assumed to be composed of identical hardware devices with fixed sets of sensors. In the envisioned Internet of Things, where every physical device will be connected, many heterogeneous devices will be present and expected to co-exist in the same network. Seamless communication between these heterogeneous devices can be seen as a prerequisite for cooperative behavior, which lays the groundwork for the intelligent automation of buildings.
Data Processing on Embedded Systems
Energy is often a scarce resource on embedded sensing systems, particularly when they are designed to be interconnected over wireless links. Local data processing mechanisms to reduce wireless traffic, e.g., by means of data compression, is a viable means to extend a node lifetimes and thus contributes to the overall operability of the entire network. We investigate (energy-)efficient and resource-aware data processing methods with a focus on their practical applicability on embedded systems.
Research projects and outputs
Tracebase is a repository of high-resolution power consumption traces from more than 30 different classes of electric appliances.
SECoM (“A Systematic Energy Information Collection Methodology for Improved Energy Analytics”) is a DFG-funded project (RE 3857/2-1), targeting to investigate the relation between the the resolution at which energy consumption data are being collected and the resulting information content.
ANTgen is a tool to create synthetic load signatures from parametric appliance models. It can be used to benchmark the performance of NILM (disaggregation) algorithms.