Distributed Averaging With Short Node Memory

This paper is explain about important process before consensus process between node sensor.

DISTRIBUTED algorithms for solving the average consensus problem have received considerable attention in the distributed signal processing and control communities recently.

This paper can explain more detail. File Size: 812 KB.

Optimization and Analysis of Distributed Averaging With Short Node Memory Click Here

Found on IEEE explore website

So, Information about this paper according my understanding is

1) Author Name : 

Boris N. Oreshkin, Mark J. Coates, and Michael G. Rabbat

2) Title Of Paper : 

Optimization and Analysis of Distributed Averaging With Short Node Memory

3) Source : 

IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 58, NO. 5, MAY 2010

4) Summary Of Reported Research :

This paper proposes on average consensus algorithm, each sensor nodes not only consider the data information from neighbors but also the data previously obtained. This is very important to obtained faster convergence time and acurate data information. 

This result applies to the general class of distributed averaging algorithms using node state prediction, and shows that, even in its simplified form and even at the theoretical level, accelerated consensus may provide considerable processing gain. 

5) Strength Of Invention:

This paper provides theoretical performance guarantees for accelerated distributed averaging algorithms using node memory. 

An important contribution of this paper is the derivation of upper bounds on the spectral radius of the accelerated consensus matrix. 

6) Weakness :

In this paper Initialization scheme of problem is complicated. Therefore Need simpler initialization schemes are the focus of ongoing investigation.

7) Limitation:

In this paper, focus on a linear predictor and derive a closed-form expression for the optimal mixing parameter one should use to combine the local prediction with the neighborhood averaging.