This report documents the development of a common interface for privacy preserving queries in pervasive computing. The interface is based on an analysis of existing solutions, and uses a simple client-server architecture, and is created with flexibility, security, accuracy, and complexity, in mind. With the interface a client has the possibility to specify a minimum cloaked area, and maximum communication allowed on a query to query basis.
Two solutions are developed under the common interface. One method based on a data-independent datastructure, called the Grid Method, and one based on a data-dependent datastructure, called the Quadtree Method. This is done to explore pros and cons of data-dependency on the server. Furthermore, it is discussed how k-anonymity can be added to the solutions.
To experiment with the two developed solutions, a test is implemented. Tests are carried out on a real world data set, a uniform distribution, and several Zipf distributions. It is found that the Grid Method performs best under most circumstances, but in the end it is indicated that this might be due to a simplifying design decision.