Distributed Weighted Least Squares Estimation with Fast Convergence in Large-scale Systems

DAMIÁN MARELLI; MINYUE FU

Florencia

Conferencia; IEEE Conference on Decision and Control (CDC); 2013

Institute of Electrical and Electronics Engineers (IEEE)

We propose a distributed method for weightedleast squares estimation. Our method is suitable for large-scalesystems, in which each node only estimates a subset of theunknown parameters. As opposed to other works, our goal isto maximize the convergence speed of the distributed algorithm.To this end, we propose a distributed method for estimating theoptimal value of certain scaling parameter on which this speeddepends. To further speed the convergence, we use a simplepreconditioning method, and we bound the difference betweenthe resulting speed, and the fastest theoretically achievableusing preconditioning. We include numerical experiments toillustrate the performance of the proposed method.