A Low-Energy Distributed Graph Signal Sampling and Reconstruction Method

Abstract

The invention discloses a low energy consumption distributed graph signal sampling reconstruction method. Existing methods have slow convergence rate and exchange more information. On the basis of the recursive least squares graph signal reconstruction algorithm, the method of the present invention constrains the consistency of the estimated values ​​of all nodes by setting a small number of bridge nodes in the network, reduces the total number of communication links in the network, further reduces communication overhead, reduces Network power consumption. The present invention is divided into an initialization part and a reconstruction part. In the initialization part, the network needs to complete tasks such as estimating the graph signal bandwidth, determining a sampling node set and a bridge node set, and initializing parameters used in reconstruction. , through the mutual communication between bridge nodes and common nodes, the update of bridge node estimators, common node estimators and dual variables is completed. The method of the invention is applicable to the scene where the observed values ​​of adjacent nodes are similar in the wireless sensor network application, can effectively reduce the energy consumption of the network, and prolong the service life of the network.

Description

Technical field [0001] The present invention belongs to the field of information technology, specifically the field of distributed signal and information processing, relates to a low energy distributed graph signal sampling reconstruction method. Background [0002] With the development of information technology and the popularization and use of the network, information processing has developed from the traditional way to the direction of network-based distributed information processing. Compared with traditional signal processing techniques, graph signal processing technology is more suitable for networked information processing, so it has gradually become one of the hot areas of current research. [0003] In practical applications, the observed values of adjacent network nodes tend to show similarities, such as the temperature difference between adjacent regions is not large, which is characterized by low frequency characteristics in the frequency domain of the graph signal. R...

AI Technical Summary

Problems solved by technology

[0006] In the existing frequency domain reconstruction method, the graph signal reconstruction algorithm based on the least mean square (Least-Mean-Squares, LMS) criterion is due to the low sampling rate of the nodes in the network, the introduction of little new information, and the slow convergence rate, and the communication of the estimated target is completed. high cost

The iteration of the graph signal reconstruction algorithm based on the Recursive-Least-Squares (RLS) criterion uses historical information, which can improve the convergence rate of the algorithm, but a lot of information needs to be exchanged in a single iteration, and the communication overhead is large.

[0007] The existing graph signal reconstruction method based on the vertex domain relies on the final estimated value of the central node and the converging node to obtain the reconstructed graph signal, which is not conducive to the distributed implementation of the algorithm

The graph signal reconstruction method based on the frequency domain is estimated to have a slow convergence rate, a large amount of information exchanged in a single iteration, high communication overhead, and high network energy consumption.

Images