Collaborative filtering based reliable data fusion optimization method in wireless sensor network

Abstract

The invention discloses a collaborative filtering based reliable data fusion optimization method in a wireless sensor network. The method adopts a hierarchical logic structure design, and respectively optimizes the network performance from two aspects such as the sensor node data fusion reliability and the link transmission reliability. Noise filtering is carried out on burst nonlinear noise data by adopting a Laplace function rather than a Gaussian function. Under a hostile environment, in order to improve the data fusion accuracy and avoid invasion of mass malicious nodes to interfere real information, the similarity of perception data is calculated by adopting an information amount vector, and weighted averaging is carried out so as to perform data fusion; and in order to enhance the transmission reliability, a link detection model is adopted to carry out prediction on a link. When the link breaks down, the cluster head sends a PROB message to inform a neighbor node of the breakdown link or a breakdown node, the neighbor node perceives important data again so as to prevent loss of the important data. The method disclosed by the invention effectively solves problems of high node deployment density, high data redundancy, limited energy and defect of being vulnerable to attacks in the wireless sensor network.

Description

technical field [0001] The invention relates to a reliable data fusion optimization method based on collaborative filtering in a wireless sensor network, which belongs to the intersection field of wireless communication technology and computer network. Background technique [0002] With the popularization of wireless sensor network applications, such as environmental monitoring, smart home, vehicle network, telemedicine, etc. The current research enthusiasm for wireless sensor networks is also intensifying, and sensors are an indispensable part of wireless sensor networks, because sensor nodes are only powered by batteries with limited energy. Under the premise of energy constraints, it needs to perceive and transmit a large amount of environmental data and forward it to the cluster head node or base station (sink node) of the cluster. The data sensed by adjacent sensor nodes is usually highly correlated, or even the same, which causes too much redundant data sensed in the ...

AI Technical Summary

Problems solved by technology

In a hostile environment, once a node is captured by the enemy, the enemy maliciously modifies the data and participates in fusion with the maximum or minimum score value, resulting in a deviation from the benchmark average value, thereby reducing the accuracy of data fusion

The traditional fusion method generally uses summation and maximum value as the reference basis for fusion, without considering the factors of malicious node intrusion and link transmission failure, which makes the network vulnerable to attacks, the accuracy of fusion data is reduced, and the redundancy is high.

Images