Channel Model And Detection Algorithm For Dual-polarized MIMO Communication System

In terrestrial wireless networking, the technology of MIMO(Multi-InputMulti-Output) can be applied to acquire great spatial multiplexing gain and diversitygain. So the idea of utilizing MIMO technology in satellite communications isgenerated. However, because of the inherent difference between terrestrial wirelesschannel and satellite wireless channel, it is a challenging subject to figure out howto apply MIMO technology in satellite communications. Abundant scatteringenvironment is needed when using MIMO technology to satisfy the condition thatthe channels are independent from each other which is difficult for satellitecommunication with LOS(Line Of Sight). Therefore, the most fundamental problemis how to generate satellite channels which are independent from each other torealize MIMO technology. Considering the spatial limitation of single satellite andthe instability of MIMO communication link caused by the unfixed locations ofsatellite points of lower and medium orbits, this article focused on the research ofsetting dual-polarized antennas in GEO(Geostationary Earth Orbit) satellites torealize MIMO communication system. Before the research of the relativetransmission characteristics of the system, what we should understand is the channeltransmission characteristics. Meanwhile, the channel characteristics can be utilizedto establish models to testify the reliability and validity of various communicationtechnologies.This article researched on single-star dual-polarized MIMO systems anddistributed satellites dual-polarized MIMO system and provided the establishmentof these systems. Additionally, statistical channel models of multi-states Markovdual-polarized MIMO over satellite channels are proposed based on the twomentioned systems. In the model establishment, this article focused on the researchof small scale fading model which introduced the effect of Doppler spectrum shift. Inaddition, proper MIMO channel model form is adopted to assure the correlations ofchannels and the stabilities of channel statistical characteristics. Then the channelmodel is simulated. Compute the cumulative probability distribution, the levelcrossing rate and the average fade duration of each sub-channel and compare withthe theoretical value to verify the theoretical correctness of the model. Meanwhile,based on the established channel model, detection formula derivation processes oftwo kinds of soft-output detection algorithms which can be defined as a preciseposterior probability MIMO detection and linear soft-output MMSE(MinimumMean Square Error) are analyzed. And on the basis of the two algorithms improvements, a modified linear soft-output algorithm with good performance andlow hardware complexity is proposed. Finally, according to the proposed channelmodel, hardware and software simulations are computed to verify the detectionperformance of the proposed algorithm.