【作者】 战金龙；

【导师】 廖桂生；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2007， 博士

【摘要】 MIMO(Multiple-Input Multiple-Output)技术充分开发空间资源,利用多个天线实现多发多收,在不增加频谱资源和天线发送功率的情况下,可以成倍地提高信道容量,满足未来无线通信中高速数据传输的需求。OFDM(Orthogonal FrequencyDivision Multiplexing)技术是多载波传输的一种,其多载波之间相互正交,可以高效地利用频谱资源。另外,OFDM将总带宽分割为若干个窄带子载波可以有效地抵抗频率选择性衰落。因此充分开发这两种技术的潜力,将二者结合起来(MIMO-OFDM)可以成为新一代移动通信核心技术的解决方案。本论文主要针对MIMO-OFDM系统,设计和研究了四种新型的发射分集方法;并结合阵列信号处理,提出了V-BLAST(Vertical Bell Labs Layered Space-Time)OFDM系统中两种稳健的检测算法。而且通过理论分析与仿真实验进行了验证。主要工作包括以下几个方面内容:1.提出了一种可以使基于准正交设计的空时分组码(QOSTBC,Quasi-Orthogonal Space-Time Block Code)获得满分集增益的准正交分组的分层空频时编码(GLSFTBC,Group Layered Space-Frequency-Time Block Coding)-OFDM发射分集方法,即QO-GLSFTBC-OFDM。该方法将4个发射天线分为两组(每组2个),输入的信号经过空频编码(SFBC,Space-Frequency Block Coding)后分成两组,然后每组的数据分别经过空时编码(STBC,Space-Time Block Coding)、OFDM调制后,由相应的发射天线发送。在接收端首先利用子载波分组进行组间干扰抑制,然后对每组分别进行译码,由于每组都是Alamouti编码,因此每组都可以获得满速率和满分集增益,将两组合并后仍可以获得满分集增益。与采用星座图旋转的QOSTBC方法相比,该方法不扩大发射天线发送符号的星座图。从理论分析和计算机仿真的结果都可以看出:该方法在保持QOSTBC满速率优点的前提下,可以同时获得满空间分集增益;而且编码和译码的过程都是基于线性处理的,计算简单。2.提出了频率选择性快衰落信道下基于线性星座图预编码(LCP,LinearConstellation Precoding)的分组的分层空频时编码OFDM发射分集方法,即LCP-GLSFTBC-OFDM。该方法的发射机结构基于四级用户码的设计:第一级基于OFDMA(Orthogonal Frequency-Division Multiple Access),主要用来进行组间干扰和码间干扰的抑制;第二级基于STBC,用来获得空间分集增益;第三级基于SFBC,用来获得时间分集增益;第四级基于LCP,用来获得频率分集增益。既克服了当发射天线数目较多时STBC设计复杂的缺点,又克服了传统分组的分层空时编码(GLSTBC,Group Layered Space-Time Block Coding)-OFDM要求信道必须为准静态衰落的弊端,而且编译码的过程都是基于线性处理的,计算简单。理论分析和计算机仿真结果也证明了该方法的有效性。3.提出了MIMO-OFDM系统中两种多用户传输方法。第一种是基于子载波分组抑制组间干扰的CDMA多用户传输方法,即分组的分层空频编码(GLSFBC,Group Layered Space-Frequency Block Coding)-OFDM-CDMA(Code-divisionmultiple-access)方法。该方法的发射机结构基于三级用户码的设计:外码(基于OFDMA)主要用来进行组间干扰和码间干扰的抑制;中间级的码(基于SFBC)用来获得空间分集增益;内码(基于CDMA)用来消除多用户干扰。第二种是基于CDMA抑制组间干扰的OFDMA多用户传输方法,即GLSFBC-CDMA-OFDMA方法。该方法的发射机结构也是基于三级用户码的设计:外码(基于OFDMA)主要用来区分多用户和消除码间干扰;中间级的码(基于CDMA)用来消除组间干扰,内码(基于SFBC)用来获得空间分集增益。这两种方法在接收端都只需要一根天线就可以同时消除组间干扰和多用户干扰,因此,大大降低了接收机的复杂度(传统的基于SVD组间干扰抑制方法和基于串行干扰相消的方法需要对每个子载波对应的信道矩阵进行相应的操作,因而复杂度较高;而且需要多个接收天线)。理论分析和计算机仿真结果证明了这两种方法的有效性。4.针对实际中信道估计存在误差,从阵列信号处理的角度提出了V-BLASTOFDM系统中两种稳健的检测算法。第一种是向信号子空间投影:将接收数据的协方差矩阵进行特征值分解,得到信号子空间,将存在估计误差的信道矢量向该子空间投影,就可以得到较为准确的信道矢量,然后利用Capon波束形成算法计算出滤波器的系数。第二种是利用信号子空间和噪声子空间相互正交:将接收数据的协方差矩阵进行特征值分解,得到噪声子空间,利用信号子空间和噪声子空间相互正交的关系优化一个代价函数,就可以得到实际信道响应的最优估计,然后利用Capon波束形成算法计算出滤波器的系数。仿真结果表明:当信道估计存在误差时,这两种方法明显优于ZF(Zero-Forcing)和MMSE(Minimum Mean SquareError)算法。更多还原

【Abstract】 MIMO (Multiple-Input Multiple-Output) technique utilizing multiple antennas to realize multiple transmission and multiple receiving, can exploit space resource adequately and can improve channel capacity without any loss in bandwidth and transmitting power. So it can meet the need of high data transmission in future wireless communications. OFDM (Orthogonal Frequency Division Multiplexing) as a kind of multi-carrier transmission can utilize spectrum resource efficiently. Furthermore, the whole bandwidth is divided into many narrow subcarriers, which can combat frequency selective fading. So the combination of MIMO and OFDM could be core solution for 4th Generation Mobile Communication. Aiming at MIMO-OFDM system, four novel transmitter diversity schemes are designed and studied, and two robust detection algorithms are proposed for V-BLAST (Vertical Bell Labs Layered Space-Time) OFDM systems based on array signal processing theory in this dissertation. Then these methods are verified by theoretical analysis and computer simulations. The primary contributions included in this dissertation can be summarized as follow:1. A novel QO-GLSFTBC (Quasi-Orthogonal Group Layered Space-Frequency-Time Block Coding)-OFDM scheme with full-rate and full-diversity is presented for improving the performance of quasi-orthogonal codes. The four transmit antennas are divided into two groups (2 of each group), after SFBC (Space-Frequency Block Coding), the input signals are divided into two groups. After STBC (Space-Time Block Coding) and OFDM modulation, the two groups are transmitted by corresponding antennas. At the receiver, using mutual orthogonality between subcarriers to suppress group interference, then the two groups are decoded, respectively. Because each group is an Alamouti coding, it can achieve full rate and full diversity. Combining the two groups can still achieve full rate and full diversity. Compared to the constellation-rotated quasi-orthogonal codes, the newly proposed scheme has the advantage of not expanding the signal constellation at each transmit antenna. Furthermore, the encoding and decoding process of the proposed method is made of linear processing and only requires simple operation. Both theoretical analysis and simulation results show the validity of the proposed method.2. A novel GLSFTBC-OFDM transmitter diversity scheme based on LCP (Linear Constellation Precoding) is proposed for wireless communications over frequency -selective fast fading channels. That is LCP-GLSFTBC-OFDM. The proposed approach is based on a four-level design of user codes: the first level which is based on OFDMA (Orthogonal Frequency-Division Multiple Access) deals with group interference and intersymbol interference (ISI), the second level which is based on STBC (Space-Time Block Coding) results in space diversity, the third level which is based on SFBC (Space-Frequency Block Coding) obtains time diversity, and the fourth level which is based on LCP results in frequency diversity. The proposed method overcomes the shortcomings of the complexity of STBC matrix when the number of transmit antennas is more than 2 and the weakness of conventional GLSTBC-OFDM, which requires quasi-static fading, furthermore, both encoding and decoding process of the proposed scheme are based on linear processing and require simple operation. Theoretical analysis and simulation results show the validity of the proposed scheme.3. Two multiuser transmitter schemes are proposed for MIMO-OFDM systems. The first one is CDMA multiuser transmitter scheme, which is based on subcarrier grouping to suppress group interference. That is GLSFBC (Group Layered Space-Frequency Block Coding)-OFDM-CDMA (Code-division multiple-access). It is based on a three-level design of user codes: the top level (based on OFDMA) deals with group interference and intersymbol interference (ISI), the middle level (based on SFBC) results in space-frequency diversity, and the lower level (based on CDMA) handles multiuser interference. The second one is OFDMA multiuser transmitter scheme, which is based on CDMA to suppress group interference. That is GLSFBC-CDMA-OFDMA. It is also based on a three-level design of user codes: the top level (based on OFDMA) deals with multiuser interference and intersymbol interference (ISI), the middle level (based on CDMA) handles group interference, and the lower level (based on SFBC) results in space-frequency diversity. Both of them only need one receive antenna to distinguish multiple users and suppress group interference simultaneously, so the complexity of the receiver decreases remarkably (the conventional group interference cancellation methods need to operate on the equivalent channel response matrix corresponding to each subcarrier, and need more than one receive antenna). Theoretical analysis and simulation results confirm the validity of the two multiuser transmitter schemes.4. Two robust detection algorithms based on array signal processing theory are proposed for V-BLAST OFDM systems with channel estimation error. The first one is projecting onto the signal subspace. The covariance matrix of receive data is eigendecomposed, and the signal subspace is obtained. Then the channel vector (with error) is projected onto the subspace, more accurate channel vector is obtained. According to Capon beamforming algorithm, the filter coefficients are calculated. The second one is according to the mutual orthogonality between the signal subspace and the noise subspace. The covariance matrix of receive data is eigendecomposed, and the noise subspace is obtained. According to the mutual orthogonality between the signal subspace and the noise subspace, a cost function is optimized, then the optimal estimation of the actual channel response is obtained. According to Capon beamforming algorithm, the filter coefficients are calculated. Simulation results show that the performance of the proposed algorithms are better than ZF (Zero-Forcing) and MMSE (Minimum Mean Square Error) algorithms under channel estimation error.更多还原