【作者】 李丹；

【导师】 冯穗力；

【作者基本信息】 华南理工大学 ， 通信与信息系统， 2010， 博士

【摘要】 IEEE 802.16协议是IEEE 802工作组制定的宽带无线通信标准,目的在于解决城域网中用户站点和核心网间“最后一英里”的高速宽带无线接入问题。IEEE 802.16e是IEEE 802.16d的移动性增强版,它在2～6GHz的特许频段内支持终端移动性,提供了一种可为用户在快速移动条件下实现高速数据传输服务的宽带无线接入解决方案。OFDM/OFDMA因其抗频率选择性衰落好,频谱利用率高等优点成为包括IEEE802.16e移动通信网络在内的下一代宽带无线通信系统的物理层核心技术。随着载频及终端移动速度的提高,移动信道在一个OFDM符号间隔内拟平稳的假设不再成立。当信道在一个码元间隔内变化时,信道时变引起的多普勒频移将破坏子载波正交性,由此造成的子载波间干扰(ICI)将严重影响系统性能。传统的OFDM时变信道估计技术在时域进行,可对一个码元间隔内的信道的时变特性进行精确估计,获得信道频域矩阵,保证在均衡时较好的去除信道时变造成的ICI。但上述研究成果均基于理想的假设条件:信道为整数倍采样信道、导频分布满足采样定理要求的理想导频模式,利用理想的导频恢复的信道时域响应无能量泄露。在实际的OFDM/OFDMA系统中,无线信道一般为非整数倍采样信道,并且系统导频通常无法满足所要求的理想模式,造成现有算法的性能严重受限。本文针对以上问题,以IEEE802.16e协议为基础,分别对实际的OFDM及OFDMA系统下的时变信道估计技术进行研究,具体研究内容及创新成果如下:(1)针对现有OFDM时变信道估计算法的局限性,结合实际的IEEE802.16e OFDM系统,本文提出了一种基于分段线性模型的OFDM频域线性时变信道估计方法,利用线性模型拟合信道频域传输函数在一个符号内的线性时变,进而估计得到信道频域矩阵,能够保证在均衡时去除信道线性时变造成的ICI。所提频域时变信道估计方法在频域处理,不受信道时域响应能量泄露的影响,适用于实际的OFDM系统。(2)针对终端移动速度增加、信道在一个符号内时变加剧,影响信道估计的精度的问题。本文提出一种在时变信道估计中利用差分编码方法消除导频子载波上来自相邻数据子载波ICI干扰的方法,以有效降低导频子载波上的ICI污染的功率水平,保证时变信道估计算法的估计精度。(3)在一个IEEE802.16e OFDM实际系统中,随着车速的增大,信道在一个符号间隔内呈现非线性时变,针对这一问题,本文提出了一种基于基扩展模型的OFDM频域非线性时变信道估计方法,利用基扩展模型拟合信道在一个符号内的非线性时变。根据信道的非线性时变程度及所需的导频开销,分别提出了利用相邻符号内导频辅助估计及利用当前符号内导频辅助估计的方法,并给出各自的适用范围。(4)本文进一步结合IEEE802.16e OFDMA协议,针对实际的OFDMA系统中用户只能利用其占用的时频单元块内的导频完成信道估计的特点,分别提出了在实际的非整数倍采样信道下,信道在一个符号内呈现线性及非线性时变时,适用于实际OFDMA系统的频域快时变信道估计方法。该方法填补了目前现有OFDMA信道估计算法无法跟踪信道在各符号内时变的技术空白,可对信道在每个符号内的线性及非线性时变进行精确拟合,保证在均衡时较好的去除信道时变造成的ICI,可为系统支持120km/h~250km/h的车载移动通信提供可靠保证。(5)最后,对以上研究成果与多入多出多天线技术的结合进行了研究讨论。得到的结论为:在移动环境下,只要各天线的导频正交,OFDM系统中已有的时变信道估计方法均可以沿用至MIMO-OFDM系统中。但是,由于多天线系统需要估计的信道参数相对单天线系统成倍增加,需要的导频开销也应随之增加。当系统导频开销一定时,多天线系统的信道估计精度相对单天线系统必然有所降低。因此,在应用中应结合具体的信道特性及系统提供的导频开销综合考虑应用多天线是否能带来效益。更多还原

【Abstract】 Broadband wireless access (BWA) based on the IEEE 802.16 standard is a promising technique for last-mile access. IEEE Std 802.16e is an enhancement to IEEE Std 802.16d to support subscriber stations moving at vehicular speeds and the operation is limited to licensed bands suitable for mobility between 2 to 6 GHz. Thereby, it specifies a system for combined fixed and mobile broadband wireless access.Orthogonal frequency division multiplexing (OFDM) is one of the most important physical layer transmission techniques in future wireless communications due to its robustness to multi-path channel and high spectrum utilization. OFDMA is a strong candidate multi-access scheme for high-data-rate mobile wireless communications, due to its flexibility on sub-carrier allocation and multi-user diversity utilization over multi-path fading channels. Hence, OFDM and OFDMA techniques are defined as the main physical layer transmission techniques in IEEE Std 802.16e.In mobile communications, the high speed motion of subscribers give rise to Doppler effects that destroy orthogonality among sub-carriers, leading to Inter-Carrier Interference (ICI) which degrades performance of the OFDM/OFDMA receiver. Conventional OFDM time varying channel estimation methods have provided promising results, but are unavailable in practical OFMD/OFDMA systems. Because the time variations of channel impulse response (CIR) are tracked in the time-domain with an ideal assumption that pilots are evenly spaced through the whole frequency band and the channel is an ideal sample-spaced channel. But concerning a practical OFDM/OFDMA system where not noly the non-sample spaced multipath delay but also the unevenly distributed pilots will lead to CIR leakage, the perforamance of conventional algorithms will decrease rapidly.Based on analysis above, this dissertation investigate the frequency domain time varying channel estimation techniques for practical IEEE 802.16e OFDM/OFDMA systems over time varying channels. The main contributions are listed as follows:( 1 ) According to practical IEEE 802.16e OFDM systems, a piece wise linear model based frequency domain time varying channel estimation method for practical OFDM system is proposed. The time variations of channel frequency transmission function are tracked by linear model using pilots in adjacent symbols. Then the channel matrix can be obtained and ICI is well migigated in equalization.( 2 ) In high mobility applications, the ICI energy increased greatly. We find that ICI on pilots make the estimated channel frequency responses at pilot sub-carriers not accurate and inevitably lead to estimation errors. To further improve the estimation precision, this paper considers the use of frequency-domain correlative coding to compress the ICI from data to pilots. The proposed method makes pilots less sensitive to ICI from data sub-carriers, especially in high Doppler environments, more favorable benefits will be received.( 3 ) With the user mobility increase, the channel varies non-linearly during one OFDM symbol period. We futher investigate a BEM based frequency domain time varying channl estimation method for practical OFDM systems. The non-linearly time variations of the channel frequency transmission function are model by BEM. According to the degree of channel variations and the pilot costs for channel estimation, two kinds of channel estimation methods are proposed respectively, which are the adjacent symbols assisted channel estimation method and the channel estimation method by using pilot clusters in the current symbol. Additionaly, their different application ranges are given.( 4 ) Conventional time varying channel estimation methods for OFDMA systems assumed that the channel is constant in a symbol period and the ICI can be ignored. It is inappropriate for high mobile environments. In our dissertation, frequency domain time varying channel estimations methods for practical IEEE802.16e OFDMA systems are proposed, which are performed in each OFDMA time-frequency blocks and the channel variariations during one OFDMA symbol period are tracked well. Hence, ICI caused by channel variations can be mitigated after equalization.( 5 ) Finally, the applicability of all above achievements in MIMO systems are discussed. Conclusions are as follows: If the pilots on different transmit antennas are orthogonal, channel estimation can be performed independently between a pair of transmit and receive antennas. Then, the above achievements will continue to apply in MIMO-OFDM/OFDMA systems. But for MIMO systems, the parameters to be estimated increase manyfold. On the premise of limited pilot overhead costs, compared to single transmit antenna systems, the eatimation performance will drop more or less. In practical applications, comprehensive considerations are necessary.更多还原