The Research of Wireless Channel Feature Extraction and Channel Discrimination

In different environments, there are some differences characteristics of the wireless channel. These differences are called "fingerprint". Extracting the "fingerprint"characteristics of different channels in different environment is very important to the devel-opment of wireless communication. This paper is focus on the problem of "fingerprint extraction". Through wireless channel signal inversion and K-means clustering and compressed sensing, establishing an adaptive clustering model. And then establish a reasonable "fingerprint"feature with the existing data and relevant physical back-ground. Use MATLAB to solve the model and verify the accuracy. Computer verification and comparison analysis show that this model can be used to identify the wireless channels with high accuracy.