On the performance of average consensus in mobile wireless sensor networks

Distributed averaging in wireless sensor networks is fundamental for advanced distributed algorithms. In this paper, we study the impact of node mobility on the well-known average consensus algorithm. More specifically, we study the MSE in mobile wireless sensor networks modeled via evolving random geometric graphs. We provide a closed-form lower bound for the MSE that shows that node mobility is beneficial. Numerical results involving various mobility models demonstrate that mobility can be traded against convergence time, transmit power, or node density.