 
Summary: The Center for Control, Dynamical Systems, and Computation
University of California at Santa Barbara
Fall 2009 Seminar Series
Presents
Performance Analysis of the Consensus Based
Distributed Estimation Algorithms
Sandro Zampieri
Dipartimento di Elettronica ed Informatica
Universita di Padova
Friday, October 23, 2009, 3:00 4:00pm HFH 1104
Abstract: Assume that a graph is given and that each node of this graph stores a real number. The average
consensus algorithm is a procedure which allows to compute the average of this set of numbers in a distributed
way, namely by exchanging information among nodes exclusively along the available links of the graph. This
problem appears in a number of different application contexts such as distributed optimization, load balancing,
data fusion for sensor networks, clock synchronization and coordinated control of mobile autonomous agents.
The performance evaluation of this algorithm depends in fact on the specific application in which this is used. In
this context we will propose some performance indices which are not typically considered in the literature. This
lecture will present then consensus based algorithms, which permit to deal with the presence of randomly time
varying networks, digital communication channels and quantized or noisy data. Finally we will show how this
technique can be applied for the solving distributed estimation and data fusion problems, clock synchronization
