Differentially private MIMO filtering for event streams and spatio-temporal monitoring

Many large-scale systems such as intelligent transportation systems, smart grids or smart buildings collect data about the activities of their users to optimize their operations. In a typical scenario, signals originate from many sensors capturing events involving these users, and several statistics of interest need to be continuously published in real-time. Moreover, in order to encourage user participation, privacy issues need to be taken into consideration. This paper considers the problem of providing differential privacy guarantees for such multi-input multi-output systems operating continuously. We show in particular how to construct various extensions of the zero-forcing equalization mechanism, which we previously proposed for single-input single-output systems. We also describe an application to privately monitoring and forecasting occupancy in a building equipped with a dense network of motion detection sensors, which is useful for example to control its HVAC system.