Sensing data error attack detection method based on random sampling consistency

Abstract

The invention relates to a sensing data error attack detection method based on random sampling consistency. The existing method needs mass of training data and is large in computation and communication expenditure. A malicious node can be effectively identified by using node location information and the signal propagation characteristic, and the method comprises the following steps: acquiring a target signal by a sensor node, collecting measurement data, and estimating the target state and identifying the malicious node, and finally comparing the target state estimation value with the preset threshold to judge whether the target is existent; the step of identifying the malicious node comprises the following steps: randomly extracting partial measurement data to estimate the target state parameter by mainly using random voting thought, and assessing the model parameter by using all measurement data, thereby finding out the optimal solution with the highest score, and identifying the node not supporting the optimal solution as the malicious node, and isolating the malicious node. The detection method disclosed by the invention is simple in operation and easy to realize; the communication expenditure is reduced since the measurement data only needs to be collected once; the applicable range is large, and the method can be used for detecting the sensing data error attack of different types.

Description

technical field [0001] The invention belongs to the field of wireless network security of communication technology, in particular to the field of distributed detection in information processing, and relates to a detection method for sensing data error attack based on random sampling consistency. Background technique [0002] Distributed detection is an important branch of distributed information processing. Distributed detection means that multiple sensor nodes in the network first measure the signal sent by the target, then fuse the observed data or local judgment results, and finally make a judgment on whether the target is in the area of ​​interest. In the distributed system structure, the identity of the sensor nodes is equal, and the load of each communication link is basically balanced, which can effectively reduce the communication cost, reduce the calculation burden, reduce the energy consumption of nodes, and avoid network congestion problems; and the distributed sy...

AI Technical Summary

Problems solved by technology

The training data of the classifier does not need to be marked in advance, but it is required that the training data cannot be mixed with wrong data, and a large amount of data is required for training

The defense method based on reputation feedback needs to first calculate the initial credibility of the node, and then evaluate the final credibility according to the credibility feedback of the diagnosis sequence and Bayesian rule, so as to adjust the classification of normal nodes and malicious nodes, but it needs Preserve historical information, and attackers may destroy the reputation mechanism

The defense method of deploying trusted nodes is to deploy trusted nodes in the network on the basis of the reputation mechanism, and make the reputation evaluation value reach a consensus among all sensor nodes through reputation propagation, which improves the stability of the reputation mechanism , but the cost of deploying trusted nodes is high, and the number and location of effective trusted nodes still need to be studied

Images