Research On Blind Identification For High Requency Channel Based On FSMC Model

Nowadays, the High Frequency (HF) channel blind identification studies mainly focus onthe linear time invariant constant of the channel model, while the true channel is time-varying,thus blind parameters identification of time-varying HF channel model is more meaningful. Ourresearch mainly focus on identifiability of the HF channel to achieve the purpose of estimatingthe channel parameters based on the variability and statistics of the HF channel model and theidentifiable conditions of the model parameters.The main work and innovation are as follows:1. The research about HF channel characteristics identification model. The actual HFchannel not only has variability but also has a Rayleigh fading. For the current channel blindidentification algrithom, linear invariant channel model application is conditionable and thevariant channel model does not has a statistical in nature, so we selected the Watterson model asa basic HF channel identification model, which is widely used in HF channel estimation. Butbecause of Watterson model is less identifiable, in order to reduce the model parameteridentification difficulty and improve the identifiability, we designed an improved finite-stateMarkov channel (FSMC) model based on Watterson model.2. The establishment of HF channel FSMC model. By studying the Rayleigh fadingcharacteristics of the HF channel, we got the statistical characteristics of the instantaneous SNRand used equal probability methods to calculate the SNR threshold. By Quantifying the HFchannel gain as a finite state Markov chain state space, we calculated the state transitionprobabilities and state step transition probability parameters, and established a HF channelFSMC model, a simplified model of Watterson. Finally, we researched the average errorprobability of the state of the model, the state average residence time and state of residence timeprobability distribution and other characteristics, and tested and verified the accuracy andapplicability of the FSMC model in describing the HF channel by experiment.3. The conditions of identifiability of HF channel based on the FSMC model. The FSMCmodel, although in the form of simplified HF channel Watterson model, but it is still a stochasticmodel. Through combined with the linear prediction algorithm and weighted Markov algorithm,we estimated the time-varying flat fading channel state, using the autoregressive model toexpress deterministic time-varying process of the channel; in addition, we study the correlationbetween each tap of frequency selective fading model channel in which contains the matchedfilter and Tap delay. Based on the FSMC model we also estimated the parameters of the relatedtap in frequency selective fading channel model. Finally, we summed up the HF channelidentifiability conditions for the above two different types of fading HF channel.4. The realization of the demonstration software of HF channel parameter extraction.Introducing the function of each module of the HF. Combined with an actual signal wedemonstrated how to extract the HF channel parameters and verified the feasibility and effectiveness of the software.