Privacy-aware Adaptive Query Processing in Dynamic Networks

Smart environments are intended to assist their users in everyday life, whether at work (Ambient Assisted Working) or at home (Ambient Assisted Living). The current situation and the actions of the user are collected through various sensors which gather a vast quantity of information. This data is stored by the system and charged up with other information, such as with the Facebook profile of the user. Through the gained knowledge, preferences, behavior patterns and future events can be calculated. Hence, the intentions and future actions of the user are derived and the smart environment reacts independently to the needs of the user. Assistance systems often collect more information than needed. In addition, the user usually has no or only a very small influence on the storage and processing of his personal data. Thus his right to informational self- determination is violated. By amending the assistance system by a data protection component, which checks the privacy claims of the user against the information needs of the system, this problem can be managed. An automated checking of different interests that may contradict or contain each other, is generally not easy to solve and requires further research. By applying an individually adapted solution to this application, the complexity of the problem can be reduced. A customized query processor, which modifies both the query and the preliminary result of the query, encourages adherence to the privacy settings of the user. Through a privacy-aware propagation of sensor and context information of the analytical tools of the assistance system not only the privacy-friendliness of the system is improved, but also its efficiency is increased. Currently, I specify a detailed scheme for describing privacy policies. In my future work, these policies will be used to modify the queries from the assistance system in order to protect the privacy of the user. Moreover, the policies will automatically adapt to new and changing data sources