On the input-output approach towards distributed estimation

International audienceIn this paper, we study distributed estimation of continuous-time, linear time-invariant systems monitored by a network of agents communicating over a graph. We assume that no agent may possess enough measurements in its neighborhood to estimate the entire state vector on its own. In this context, we provide a networked Kalman-type estimator that combines prediction and innovation with information fusion among the neighboring agents and consider an approach based on designing static estimator gains. The main contribution of this paper is to analyze the estimation error using the notions of dissipativity and the input-output approach, which enable us to formulate stability and performance arguments as quasiconvex optimization problems involving linear matrix inequalities. We show that the resulting estimation error is stable and further ensures a given level of performance regarding noise rejection. Simulations illustrate the concepts described in this paper