摘要
现行的自动识别方法无法准确识别某些复杂的调制方式.针对此问题,文章提出基于人工智能模型的通信信号自动调制识别方法.首先,在构建通信信号调制模型时,采用人工智能来判定调制方式,并给出相应的调制参数.其次,采用maxout函数替代原有的ReLU函数,同时增加卷积层的数量,减少池化层的层数,提取通信信号的特征参数,从而构建适合人工智能模型训练的特征矢量.最后,结合自适应调整的识别阈值进行通信信号的初始自动调制识别,并通过相似度进一步优化调制识别方法.通过测试数据集,验证该方法的有效性和准确性.实验结果显示,相较于文献方法,基于人工智能的通信信号自动调制识别方法具有更高的精准度.
Abstract
The current automatic recognition methods cannot fully and accurately identify certain complex modulation modes.This article proposes an automatic modulation recognition method for communication signals based on artificial intelligence models to address this issue.Firstly,when constructing a communication signal modulation model,artificial intelligence is used to determine the modulation method and provide corresponding modulation parameters.Next,the maxout function is used to replace the original ReLU function,while increasing the number of convolutional layers and reducing the number of pooling layers to extract the feature parameters of the communication signal,thereby constructing a feature vector suitable for artificial intelligence model training.Finally,this article combines the adaptive recognition threshold for initial automatic modulation recognition of communication signals,and further optimizes the modulation recognition method through similarity.The effectiveness and accuracy of this method were verified through testing the dataset.The experimental results show that compared to literature methods,the communication signal automatic modulation recognition method based on artificial intelligence has higher accuracy.
NETL
NSTL
万方数据