Automatic Device Function Extraction of IoT Devices
IoT devices need to be described according to a common vocabulary to enable machine-to-machine interaction. However, currently these devices are described manually by a device manufactorer. The aim of this master thesis is to automatically or semi-automaticcally identify the capabilities of IoT devices (what are the different actions the device support e.g., turn lamp on, turn lamp off, increase lamp brightness, etc), the precondition for executing a device action (e.g., the lamp should be off if we want to turn it on) and the possible effect a device action will have (e.g., turning the lamp on will result in an increase brightness). The current solutions in the literature assume that this information is provided by an expert. However, this is not feasible because:
- Time consuming process
- Information Redundancy
Therefore, a better, efficient, and time-saving process needs to be provided. One solution could be to automatically extract some information from an IoT device API. Another approach could be to use knowledge graphs like ConceptNet to identify some information.
Interested users can contact me for more information.