MAC information identification method, device, and equipment based on GRU network, and storage medium

Abstract

The invention provides an MAC information identification method, device and equipment based on a GRU network, and a storage medium. The method comprises the steps of obtaining an MAC information identification network model and a channel type identification network model under different channels through training; receiving a to-be-identified signal; calculating the signal power of the to-be-identified signal; selecting a plurality of sampling points in the signal power as a channel identification observation vector and an MAC information identification observation vector; identifying the channel identification observation vector by using a channel type identification network model to obtain a channel type identification result; and according to a channel type identification result, selecting an MAC information identification network model to identify an MAC information identification observation vector so as to judge the number of main users of the signal to be identified and the MAC protocol type. According to the method, the time domain correlation of MAC protocol signals is considered, the characteristic that a GRU network is good at processing a time sequence is utilized, the channel type can be recognized firstly, and then the MAC protocol type and the number of main users are recognized under the condition that the channel type is determined.

Description

technical field [0001] The present invention relates to the field of cognitive radio, in particular to a GRU network-based MAC information identification method, device, equipment and storage medium. Background technique [0002] Cognitive radio, as a spectrum sharing technology with learning ability, alleviates the situation of spectrum shortage. Considering that the MAC protocol sequence in the cognitive radio system contains rich semantics under different channels, multi-label classification of the primary user signal can better reflect the true meaning of the primary user signal sequence and obtain more prior knowledge of the primary user. Information, so that secondary users can adjust network parameters more accurately and reduce interference to primary users. At present, the identification work of the MAC protocol mainly focuses on the research of single-label classification, and only one label of the primary user signal of the MAC protocol is identified, which has c...

AI Technical Summary

Problems solved by technology

Different from single-label classification tasks, there are mutual exclusion, co-occurrence and other relationships between multi-label examples, which also makes it more challenging to deal with multi-label classification tasks

Secondly, the increase in labeling cost caused by too many labels also poses a challenge to the training model

Images