The feature extraction and matching of time-varying fast fading channel

Forthe issues of feature extraction and scene recognitionin the complexwireless transmission channel, this paper presents a feature recognition based on a "fingerprint library" algorithm. The algorithm is fully tap the different scenarios, differentiated features within the different regions of the radio channel, for different channel parameters were extracted using the improved KNN algorithm to classify the channel scene. Simulation analysis and experimental data show that the algorithm solves the channel characteristics problemsof wireless channel parameter extraction and channel scene recognition.