PhD Thesis (draft in Chinese): A Joint Transceiver Design Framework for Iterative Interference Cancelation Systems

[GoogleCodeExporter]

This is a draft version of my PhD thesis

Original issue reported on code.google.com by wayne.ha...@gmail.com on 31 Aug 2013 at 4:40

Attachments:

• thesis_main.pdf

[GoogleCodeExporter]

[deleted comment]

[GoogleCodeExporter]

Abstract:

Physical layer transmit/receive signal processing is one of the key techniques 
that guarantee the link performance of a wireless communication system. Joint 
transceiver design optimization is the promising scheme that utilizes spatial 
degrees-of-freedom (DoF) of multi-input multi-output (MIMO) wireless channels. 
Conventional joint transceiver design framework has already confirmed the 
unique performance advantages of the non-orthogonal transmit/receive structure, 
in terms of a variety of Schur-convex link performance measures, e.g., average 
mean-square-error (MSE) or symbol-error-rate (SER), etc. To mitigate the 
interference induced by the non-orthogonal structure, advanced signal 
processing techniques have been extensively adopted at the  receiver, where the 
iterative interference cancelation has emerged as a powerful and attractive 
candidate.

Despite the maturity of both fields in recent years, the gap between 
conventional joint transceiver design framework and iterative interference 
cancelation techniques is still wide-open, which provides enormous challenges 
and opportunities for future high performance wireless communication systems. 
This dissertation   thus considers the fundamental problem of joint transceiver 
design optimization for iterative interference cancelation systems. After an 
extensive overview of MIMO channels and conventional joint design framework, 
both ideal single/multiple-user iterative transceiver and more practical robust 
iterative transceiver against channel estimation errors are examined.

First, evolutional analysis and joint transceiver design optimization are 
studied for single-user iterative cancelation system. A dynamic statistical 
analysis approach is taken to devise a novel semi-analytical tool that 
precisely tracks and predicts the performance evolution along the iterative 
cancelation process. This evolutional analysis also provides convergence 
constraints that guarantees a successful iteration to the target  link 
performance. A joint QoS/QoC constrained power allocation problem is thus 
established, which significantly distinguishes from conventional joint 
transceiver design framework. By revealing the underlying sparsity in the 
mathematical structure of the problem, a highly efficient IPM solver is 
derived, which not only saves much of the computational cost, but also provides 
significant performance enhancement over conventional schemes.

Then, evolutional analysis and joint transceiver design optimization are 
studied for multi-user iterative cancelation system. In particular, a 
multi-access system with an iterative multiuser detector (MUD) and multiple 
independent transmitters are considered. Due to multi-user coupling, an 
alternating projection approach is taken to project the high-dimensional 
statistics into a lower sub-space, thus facilitate the evolutional analysis in 
an economical way. Based on the multi-user convergence constraints introduced 
by the evolutional analysis, a joint QoS/QoC constrained multi-user power 
allocation problem is obtained, which is non-convex and very complicated. By 
revealing the underlying partial convexity and sparsity in the mathematical 
structure of the problem, it is transformed into a generalized Nash equilibrium 
problem (GNEP) that consists of a series of  convex sub-problems. Two types of 
highly efficient iterative IPM solver are thus developed, which guarantee 
significant performance advantages over conventional schemes.

Finally, the more realistic situation of imperfect channel knowledge due to 
channel estimation errors is considered. The previous results on joint 
transceiver design for iterative cancelation systems are then extended to 
obtain robust designs. In particular, worst-case and stochastic robust design 
constraints and corresponding robust iterative receiver structures are first 
developed, according to deterministic and stochastic error modeling, 
respectively. Worst-case and stochastic robust joint transceiver design 
frameworks are then established for both single-user and multi-user iterative 
interference cancelation scenarios. Since the robust design shares the same 
basic structure with the ideal design, the aforementioned analysis tools and 
efficient problem solvers are readily applied here, which eventually leads to a 
unified general framework for single/multi-user joint transceiver design with 
iterative interference cancelation and a variety kinds of channel knowledge.

Original comment by wayne.ha...@gmail.com on 31 Aug 2013 at 4:52

[GoogleCodeExporter]

摘要：

无线通信系统物理层发射与接收信号处理技术是决定链路性��
�的核心因素之一，而收发机联合优化设计则是合理利用多输�
��多输出(multiple-input multiple-output, 
MIMO)信道所提供空域资源(spatial 
resources)的必由之路。收发机联合优化设计的经典理论表明，�
��正交收发结构之于舒尔凸类链路性能函数(包括平均均方误��
�或误符号率等常见性能指标)具有明显优势。为抑制由非正交
结构引入的干扰，接收端往往采用高级信号处理技术，其中��
�以迭代干扰抵消技术最具性能优势。

与传统收发机联合优化设计理论以及迭代干扰抵消技术近年��
�蓬勃发展日臻完善形成鲜明对比的是，基于迭代干扰抵消技�
��的收发机联合优化设计理论研究目前尚处于起步阶段，充满
机遇与挑战。因此，在简要概述广义MIMO信道并梳理和总结传�
��收发机联合优化设计的经典理论和方法后，本论文从单用户
通信系统、多用户通信系统以及存在信道估计误差的单/多用�
��通信系统着手，研究迭代收发机联合优化设计的统一理论框
架与高效信号处理算法。

首先，本论文研究了单用户迭代收发机的性能演进分析与联��
�优化设计问题。针对迭代接收性能随迭代过程动态演进这一�
��点，我们提出了一套性能演进分析工具，能够精确跟踪和预
测单用户迭代收发机的性能演进过程，从而不但为其性能评��
�提供坚实理论依据，亦为确保迭代向目标链路性能收敛指出�
��化设计途径。在此基础上，我们建立起高度结构化的受QoS和
QoC联合约束的迭代收发机最小化功率分配模型，并设计一种��
�用IPM算法，通过有效利用上述优化问题数学结构所蕴含独特�
��疏性，实现上述凸优化问题的高效求解。

其次，本论文研究了多用户迭代收发机的性能演进分析与联��
�优化设计问题。针对多用户系统存在用户间相互耦合这一特�
��，我们提出了一套基于交替演进映射的多用户性能演进分析
工具，将任意用户迭代检测高维统计量映射至一低维子空间��
�从而获得多用户性能演进分析在复杂度与精确性之间的良好�
��中。在此基础上，我们建立起受QoS和QoC联合约束的多用户迭
代收发机最小化功率分配模型。针对这一复杂非凸优化问题��
�我们进一步发掘其数学结构所蕴含局部凸性与独特稀疏性，�
��而将其转化为包含一系列凸优化子问题的广义纳什均衡问题
。在此基础上，我们设计一类迭代IPM算法实现多用户联合优��
�设计问题的高效求解。

在上述迭代收发机联合优化设计基础上，本论文进一步研究��
�针对信道估计误差的鲁棒性迭代收发机联合优化设计问题。�
��过深入分析确定性误差模型和随机性误差模型，我们分别推
导出针对最差情况和基于统计意义的鲁棒性设计约束条件以��
�相应的鲁棒性迭代接收机结构。在此基础上，我们分别提出�
��适用于确定性误差模型的针对最差情况的迭代收发机鲁棒性
联合优化分析与设计框架以及适用于随机性误差模型的基于��
�计意义的迭代收发机鲁棒性联合优化分析与设计框架。鉴于�
��述两类鲁棒性设计与前述理想信道条件下的设计具有相同的
数学结构，故前述一系列演进分析工具及相应高效优化算法��
�得以沿用，从而最终构建起一整套适用于多种不同信道估计�
��件的单/多用户迭代收发机统一优化设计框架。

Original comment by wayne.ha...@gmail.com on 31 Aug 2013 at 4:52

[GoogleCodeExporter]

An update of draft thesis

Original comment by wayne.ha...@gmail.com on 10 Sep 2013 at 3:05

Attachments:

• thesis_main.pdf

[GoogleCodeExporter]

An update of draft thesis

Original comment by wayne.ha...@gmail.com on 28 Oct 2013 at 7:27

Attachments:

• thesis_main.pdf