Differentially-Private Dynamic Average Consensus

We address differential privacy for dynamic average consensus. Not only is dynamic average consensus widely used in cooperative control and distributed tracking, it is also a fundamental building block in numerous distributed computation algorithms such as multi-agent optimization and distributed Nash equilibrium seeking. By co-designing the dynamic average consensus mechanism and the differential-privacy noise injection mechanism, we propose the first dynamic average consensus algorithm that can ensure both provable convergence to the exact average reference signal and rigorous $\epsilon$-differential privacy, even when the number of iterations tends to infinity. Given that dynamic average consensus includes the static average consensus as a special case, the approach can also be used to ensure rigorous $\epsilon$-differential privacy in static average consensus while maintaining accurate consensus result. To our knowledge, ensuring both provable convergence and rigorous $\epsilon$-differential privacy (even for infinite number of iterations) has not been achieved before in average consensus algorithms. Numerical simulation results confirm the effectiveness of the proposed approach.