MIMO-OFDM系统信道估计与信号检测联合方法研究

发布时间：2018-01-17 13:02

本文关键词：MIMO-OFDM系统信道估计与信号检测联合方法研究　出处：《哈尔滨工程大学》2014年博士论文　论文类型：学位论文

【摘要】：在MIMO-OFDM系统中,信号检测算法的性能和复杂度对整个通信系统的可靠性和实时性有很大的影响。一般具有优异检测性能的信号检测算法复杂度较高不适用于实际应用。因此研究具有最优检测性能且复杂度适中的信号检测算法是MIMO-OFDM接收系统设计的关键。另外,在系统的接收端只有对信道的精准估计才能有效地消除信道对发送信号的干扰,因此信道估计是MIMO-OFDM系统接收的重要技术之一。在MIMO-OFDM系统中,信道估计和信号检测一般是相互独立的。由于信道状态信息的获得只取决于训练序列,若信道估计精度过低,必然导致系统的误码率急剧升高。同时,大多数信号检测算法均是针对没有信道编码的无线通信系统,检测一般采用硬判决,算法没有有效的利用信道译码输出的有用信息。因此将信道估计、信号检测、信道译码通过迭代技术联合起来可有效地提高MIMO-OFDM系统的整体性能。论文针对MIMO-OFDM系统中的联合信道估计、信号检测和信道译码技术进行了深入的研究,主要研究成果如下:首先,研究了 MIMO-OFDM系统结构、最优导频设计方法以及辅助导频下的信道估计算法,给出了相应的理论分析。在此基础上针对多输入多输出正交频分复用系统的盲信道估计算法存在收敛速度慢以及使用奇异值分解复杂度高等问题,提出了一种基于改进子空间的MIMO-OFDM半盲信道估计算法。算法首先利用快速数据投影方法替代奇异值分解跟踪噪声子空间来估计信道状态信息,然后利用训练序列对信道进行校正,以解决一般盲信道估计算法存在模糊矩阵问题,提高了信道估计算法的精度。仿真结果表明该方案具有较快的收敛速度且降低了运算复杂度,是一种实用的半盲信道估计方法。其次,研究了分层空时编码、经典的MIMO系统信号检测技术,并对其进行相应的理论与仿真分析。在此基础上将量子搜索算法引入检测系统并对其进行改进,实现了一种理论上复杂度较低的改进量子搜索信号检测算法,仿真结果表明该算法能够逼近最大似然检测算法的检测性能,且系统复杂度有了明显改善。针对球形译码算法初始半径的选取和最终输出判决问题,提出了一种适合高信噪比的改进球形译码检测算法。针对基本粒子群优化信号检测算法易于陷入局部极值和收敛精度较低的问题,提出了一种适合低信噪比的改进粒子群优化信号检测算法。算法结合遗传算法的杂交技术和极值扰动机制,对基本粒子群优化算法进行改进。仿真结果表明该算法有效地提高了系统的信号检测性能,有较强的全局搜索能力,是一种实用的信号检测方法。然后,对期望最大化算法及其扩展进行理论研究。在此基础上提出了基于广义空间迭代期望最大化的时域和频域的信道估计算法,并结合信号检测理论,针对采用期望最大化算法存在收敛速度慢等缺陷,提出了一种基于频域的SAGE-ISD联合信道估计检测算法。方案主要采用了线性最小均方误差信道估计对信道进行初估计,并利用联合迭代技术结合广义空间迭代期望最大化算法和改进球形译码检测算法对信道进行估计校正和信号检测,从而提高了系统的可靠性。针对时域上最大似然检测算法难于硬件实现以及传统的信道估计性能较差等缺陷,提出了一种基于时域的SAGE-IPSO联合信道估计检测算法。该方案使用训练序列进行信道初估计,利用广义空间迭代期望最大化算法对信道估计值进行校正,并结合改进的粒子群优化算法完成对信号的迭代检测,使系统性能得到改善。仿真结果表明两种方案能以较少的迭代次数完成信道的精准估计和数据检测,检测性能与理想信道估计下的最大似然检测算法相接近。最后,对信道编译码算法和软输入软输出信号检测进行理论研究。在此基础上针对传统的MIMO-OFDM接收机中信道译码、信道估计和信号检测各自独立工作带来的误码率较高的问题,提出了一种改进的迭代联合接收方案,该方案充分利用译码器提供的软信息进行联合信道估计与信号检测,仿真结果表明所提方案显著提高了系统的性能。
[Abstract]:In the MIMO-OFDM system, signal detection performance and reliability of the algorithm complexity of the whole communication system and real time has a great influence. Signal detection algorithm has excellent performance and high complexity of detection is not suitable for practical application. Therefore research on signal detection algorithm has optimal detection performance and moderate complexity is the key to the design of receiving system MIMO-OFDM. In addition, at the receiving end of the system only on channel estimation precision can effectively eliminate the interference of the transmitted signal channel, so channel estimation is an important technology of the MIMO-OFDM system. A receiver in the MIMO-OFDM system, channel estimation and signal detection is generally independent of each other. Since the channel state information is obtained only in the training sequence, if the channel estimation accuracy is too low, will inevitably lead to the error rate increased sharply. At the same time, the majority of signal detection algorithms are In view of the wireless communication system without channel encoding, usually adopt the hard decision algorithm, there is no effective use of the useful information of channel decoding output. So channel estimation, signal detection and channel decoding through the combined iterative technique can effectively improve the overall performance of the MIMO-OFDM system. The thesis focuses on the joint channel estimation in MIMO-OFDM system, signal detection and the channel decoding technology is studied, the main research results are as follows: firstly, research the MIMO-OFDM system structure, the optimal pilot design method and auxiliary guide channel frequency estimation algorithm, gives the corresponding theoretical analysis. Blind channel based on multiple input multiple output orthogonal frequency division multiplexing system estimation algorithms the slow convergence speed and the use of singular value decomposition complexity problems, proposes a semi blind subspace based on improved MIMO-OFDM Channel estimation algorithm. The algorithm firstly fast data projection method instead of singular value decomposition of tracking noise subspace to estimate the channel state information, then the channel was corrected with the training sequence, blind channel estimation algorithm to solve general fuzzy matrix problem, improve the accuracy of channel estimation algorithm. The simulation results show that this scheme has fast convergence speed and reduce the computational complexity, is a semi blind channel estimation method is practical. Secondly, study the layered space-time encoding, MIMO signal detection classic, and carried on the theory and the corresponding simulation analysis on it. On the basis of quantum search algorithm is introduced to detect system and improvement of the implementation of the an improved quantum theory with low complexity signal detection algorithm, the simulation results show that the performance of the algorithm is able to approximate the maximum likelihood detection algorithm , and the complexity of the system is improved. According to the selection of the initial radius of the sphere decoding algorithm and the final output decision problem, proposed an improved sphere decoding for high signal-to-noise ratio detection algorithm. According to the basic particle swarm optimization signal detection algorithm is easy to fall into local minima and the convergence precision is low problem, put forward a suitable for low SNR and improved particle swarm optimization algorithm. The signal detection algorithm based on genetic algorithms and hybrid technology and extreme disturbance mechanism, the basic particle swarm optimization algorithm. The simulation results show that the algorithm can effectively improve the performance of signal detection system, has a strong global search ability, is a kind of practical method of signal detection. Then, the expectation maximization algorithm and its extension theory research. Based on the generalized space iteration based on expectation maximization in time domain and frequency domain channel Channel estimation algorithm, and combined with the signal detection theory, aiming at using the expectation maximization algorithm has the defects of slow convergence, this paper proposes a detection algorithm of joint channel estimation based on SAGE-ISD frequency domain. The scheme mainly adopts linear minimum mean square error channel estimation of channel estimation, and the use of joint iterative technique combined with generalized iterative expectation the maximum algorithm and improved sphere decoding algorithm for the estimation of signal detection and correction of the channel, so as to improve the reliability of the system. According to the time domain maximum likelihood detection algorithm is difficult to realize, the traditional channel estimation performance and other defects, this paper proposes a detection algorithm of joint channel estimation based on SAGE-IPSO time domain. This scheme using training sequence channel estimation, using the generalized iterative expectation maximization algorithm to correct the estimated value of the channel, and the knot With the improved particle swarm optimization algorithm to complete the iterative detection of the signal, the system performance is improved. The simulation results show that the two scheme can complete the accurate channel estimation and data detection with fewer iterations, the detection performance and ideal channel estimation is close to the maximum likelihood detection algorithm. Finally, the algorithm and channel coding soft input soft output signal detection theory research. Based on the traditional MIMO-OFDM receiver channel decoding, bring channel estimation and signal detection work independently with high bit error rate problem, proposed joint receiver scheme is an improved iterative joint channel estimation and signal detection, the scheme makes full use of soft information to provide the decoder the simulation results show that the proposed scheme significantly improves the performance of the system.

【学位授予单位】：哈尔滨工程大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN919.3

【相似文献】