MIMO Transmission Techniques For GEO Mobile Satellite Communications

Geostationary earth orbit (GEO) mobile satellite communications (MSC), the use of artificial GEO satellites to provide communication links between various locations on earth, play a vital role in the global telecommunications. The development of multiple-beam antenna t,hat has been recognized as an effective means to increase the capacity as it can support the transmissions of independent information streams to multiple users at the same spectrum band, shows a bright prospect of high rate GEO MSC. However, in spite of the much improved capacity, inter-beam-interference (IBI) still lead to detection errors.It is reasonable to regard the multibean channel with full frequency reuse as a multi-input multi-output, (MIMO) channel and thus we can employ MIMO transmission techniques to reduce IBI. We investigate some MIMO transmission techniques for multibeam GEO MSC with full frequency reuse in this thesis.Firstly, we propose a closed-form ergodic capacity upper bound for multibeam GEO MSC MIMO channel that reveal the relationship between the ergodic capacity and some statistical parameters of the channel. The ergodic capacity upper bound can be employed to estimate the possible transmission sum rate. Hence, we propose a user scheduling algorithm based on the closed-form ergodic capacity upper bound to optimize the trans-mission sum rate. In most case, the more users there are the greater ergodic capacity will be. However, the ergodic capacity does not increase linearly with the users. Thus,the average transmitting rate of each user may decrease with the increasing of users. The proposed user scheduling is a useful method to guarantee the transmission sum rate and average transmitting rate of each user both.Then, we study the statistical property of dual-polarized multibeam GEO MSC and propose a statistical channel modeling for it. Our work shows that the dual-polarized inultibeam antennas have much greater capacity than traditional ultibeam antennas.Moreover, we introduce the proportional fair (PF) scheduling into the dual-polarized multibeam GEO MSC to reduce its IBI. PF scheduling can not only reduce IBI, but also guarantee all users share communication resources fairly. The main problem lays in that the PF scheduling requires estimating the possible ergodic capacity, but it is very difficult. We propose a solution for this problem via the ergodic capacity upper bound as well.Subsequently, we propose a distributed joint detection and decoding method for the uplink of multibeam GEO MSC based on approximate message passing (AMP) algorithm.Beside user scheduling, joint detection and decoding in forward error correction (FEC)coded MIMO systems is also a popular technique to reduce the IBI and so the bit error rate (BER). However, the computational complexity of traditional minimum mean-square error (MMSE) joint detection and decoding that yells near capacity performance, which increase exponentially with the system size limits much its use for practice. In this pa-per, we propose a novel joint detection and decoding method based on AMP algorithm.The AMP algorithm can serve as a detector for MIMO channel, which has much lower complexity than the majority of traditional detectors. We give a solution to hybridize the AMP algorithm detector with LDPC soft-in-soft-out decoder through feeding back the in-formation of the LDPC soft-in-soft-out decoder, which can reduces detection errors much,and so reduces the decoding BER much. The proposed method no only has much lower complexity, but also has lower BER than traditional MMSE joint detection and decoding methods.At last, we propose a joint distributed precoding method for the downlink of multi-beam GEO MSC. The downlink of nultibeam GEO MSC with multibeam antenna on board the satellite serving multiple user terminals (UTs) each with a single antenna is also an MIMO channel, and it also suffers form IBI. This motivates lots of research interests in multiple-beam cooperation to turn the unwanted IBI into useful energies. The down-link cannot be detected jointly, and so zero-forcing (ZF) precoding attracts considerable attentions for its near-capacity performance and moderate complexity. Regularized ZF precoding (RZF), as a generalization of ZF precoding, is one of the most popular IBI cancelation technologies. The implementation of centralized RZF precoding is still a big challenge for MSC systems of large size mainly due to the heavy computation in matrix inversion as well as the requirement on full CSI at the processing center. To overcome these issues, there are many research in distributed precoding methods in recent years.Unlike centralized RZF where the precoded signals are generated together, distributed methods produce precoded signals through a group of computing units that work inde-pendently with the signal and CSI information of their local BMs or nearby BMs only.We introduce the distributed AMP RZF precoding scheme into the multiple-beam MSC downlink. Moreover, we propose an interference cancellation inessage passing (IC-MP) al-gorithm that combines interference cancellation into AMP The IC-MP algorithm not only converges faster than the AMP algorithm, but also has lower mean square error (MSE)and higher average achievable sum-rate in the low signal-to-noise-radio (SNR) regime,whereas the AMP algorithm has better performance in the high SNR regime. Hence, we further hybridize the AMP and IC-MP algorithms and propose a hybrid message passing algorithm to take the advantages of both.