Parameter Estimation Assisted Link Adaptation

In recent years,scattering channel has been widely used in military field because of its good stability and anti-interference performance.However,the special fading of scattering channel will reduce the spectrum efficiency,so this paper proposes a link adaptive scheme to solve this problem.In order to better realize the link adaptation scheme under the scattering channel,this paper analyzes the characteristics of the scattering channel,and proposes to use TDL(Tapped Delay Line)to describe the scattering channel model.Because the fading degree is different at different times,neither Rayleigh model nor Rice model can accurately describe its fading characteristics,so this paper uses Nakagami-m fading model to describe its fading characteristics.When the m parameter is large,it means that the fading degree is small,while the m parameter is small,it means that the fading is serious.On this basis,a link adaptation scheme based on parameter estimation is proposed in this paper.The scheme focuses on two aspects:one is the estimation of fading parameter m,and the other is the adaptive scheme based on fading parameters.The scattering channel has a long transmission distance and a large path loss,so the general communication is in a low SNR environment.In this paper,the m parameter estimation algorithm under low SNR is proposed.This paper first introduces the traditional moment estimation algorithm and maximum likelihood algorithm,considering the practicality of the parameter algorithm,this paper focuses on the moment estimation algorithm.Secondly,such algorithms have high requirements on the channel environment and can only be applied to high SNR,which does not meet the requirements of scattering channel.Therefore,this paper also explains mathematically that the algorithms mentioned in the literature are biased under low SNR and cannot be applied to the m parameter estimation in this paper.Finally,the additional term of the influence moment estimation algorithm under low SNR is derived from the point of view of mathematics and mathematical modeling is carried out.On this basis,a modified term is proposed to reduce the influence of the additional term.The final simulation results show that the algorithm proposed in this paper is far better than the traditional algorithm in both accuracy and stability.On the basis of solving the problem of m parameter estimation,this paper also focuses on the link adaptation scheme based on parameter estimation in scattering channel.This part first introduces the traditional SNR based look-up table method and BayesLA algorithm,which shows that BayesLA algorithm can make better use of the threshold margin compared with the look-up table method,so as to improve the throughput.Then,aiming at the shortcomings of these two one-dimensional characteristic parameter algorithms,the significance of fading parameter estimation is given.Then,this paper focuses on two link adaptation schemes for parameter estimation,which are based on frequency domain channel fading parameters and Nakagami-m parameters.Based on the combination of frequency domain channel fading scheme and KNN(K-Nearest Neighbor),RF(Random Forest)and other classification algorithms,this paper proposes a new fading characteristic parameter to characterize channel fading,obtains the classification model through off-line learning,and then obtains the best MCS(Modulation and Coding Scheme);The scheme based on Nakagami-m parameter directly selects the best MCS according to m parameter and SNR,which is relatively simple and does not need offline learning.Simulation results show that the two parameter estimation algorithms greatly improve the system throughput compared with the traditional algorithms.