Research On Link Adaptive Transmission Technologies For Troposcatter Communication

Troposcatter channel is widely used in ultra-short wave over-the-horizon wireless communication for its strong anti-destruction,anti-interference and obstacle-jump abilities.However,it is challenging to use the troposcatter channel to carry out ultra-long distance communication due to the high transmission path loss,the low SNR environment,and the existence of time and frequency dual selective fading.According to the characteristics of the channel,the link adaptive transmission technologies for troposcatter communication are studied,including channel quality measurement,multiple rates of MCS transmission and switching control scheme.The characteristics of troposcatter channel are analyzed from the perspective of large-scale fading,small-scale fading and three-dimension image of time-frequency amplitude of the channel.And then combined with the classical troposcatter channel,the simulation channel model is established.A Single Carrier Interleaved Frequency Division Multiplexing?SC-IFDM?communication system with time-frequency two-dimension joint framework is selected to effectively overcome the complex channel characteristics,time-frequency dual selective fading of the channel.In channel quality measurement,low SNR estimation and link quality metrics?LQM?mapping algorithms are studied.By using the unique structure of SC-IFDM,a new non data-aided SNR estimation algorithm is proposed.Compared with the previous algorithms,the proposed algorithm significantly improves the estimation range and accuracy of low SNR,and the credit range of estimated SNR extends to about-18dB,the Normalized Mean Square Error?NMSE?is blow 10^-1.Moreover,the proposed algorithm requires non data-aided,which can effectively save the cost of pilot symbol of the system.The accurate estimation of low SNR lays a foundation for the subsequent research on LQM mapping algorithms and parameter switching control schemes.Then,two common LQM mapping algorithms are studied.One is the direct mapping function algorithms,the EESM and LESM mapping algorithms are introduced,the other is the Mutual Information Effective Mapping?MIESM?algorithm,which mainly introduces the Mean Mutual Information per Bit?MMIB?algorithm.The one dimension effective SINR is obtained by the mapping algorithms,used in the traditional look-up table switching control scheme.Finally,the feedback strategy of channel quality is studied to improve the feedback efficiency.In link adaptive transmission,multiple rates MCS transmission and switching control schemes are studied.To make the transmission of troposcatter communication system robust,the transmission schemes of channel coding and rate matching,modem and the extraction of the soft information,frequency diversity combining are optimized,respectively.By combining the code rate,modulation mode and diversity order,a MCS transmission scheme with 11 different information rates is designed.The transmission scheme meets the business requirements at different rates.The range of the transmission rate for a single user is100kbps¹.25Mbps.Aiming at the inaccuracy of link feedback based on one-dimensional LQM,the machine learning method is applied to link adaptive switching control.The switching control scheme of MCS is dynamically adjusted by training sample data,to adapt to the change of channel environment.There are two machine learning methods,SVM based on classification idea and Q learning based on maximum return function,are studied to optimize the switching control strategy of MCS,and it improves throughput performance of link communication system.By adopting the machine learning method,the high-dimensional channel quality information is retained,which makes the switching control of MCS more accurate,and it effectively improves the throughput of the troposcatter communication system.