Differential privacy is a popular notion of privacy aimed at quantifying
and controlling the risk for an individual to participate in a statistical
database. We discuss scenarios where one would like to provide such 
privacy guarantees for dynamical systems, e.g., monitoring and control 
systems such as smart grids or intelligent transportation systems. We 
then discuss how system theoretic and signal processing tools allow 
us to design differentially private filters for various types of dynamic 
data. We remark that if statistical models of the users' signals are available,
exploiting them can lead to significant performance improvements for 
differentially private mechanisms, without sacrificing the privacy guarantees.