摘 要
隨著現(xiàn)代無線通信技術(shù)的快速發(fā)展，人們對高速數(shù)據(jù)通信服務(wù)的需求日益增長，常規(guī)單天線收發(fā)通信系統(tǒng)的容量性能已經(jīng)遠遠不能滿足實際應用的需求，通信系統(tǒng)的可靠性也有待進一步提高。MIMO技術(shù)可以有效利用多徑效應，在有限帶寬內(nèi)提高了傳輸速率和通信質(zhì)量，成為新一代無線通信系統(tǒng)中的關(guān)鍵技術(shù)之一。但由于MIMO系統(tǒng)需要配置多個RF鏈路，大幅度增加了系統(tǒng)硬件成本和信號處理的復雜度，這個問題極大的限制了MIMO技術(shù)的發(fā)展和推廣。天線選擇技術(shù)通過選擇較優(yōu)的天線子集進行收發(fā)，以較小的性能損失換取硬件成本和處理復雜度的大幅度降低，成為MIMO無線通信領(lǐng)域的一個研究熱點。
對于現(xiàn)有MIMO系統(tǒng)天線選擇算法而言，窮盡搜索算法能夠達到最優(yōu)的性能，但包含較多矩陣運算，計算復雜度較高；為降低算法復雜度，許多次優(yōu)天線選擇算法應運而生，但這些算法大多造成了較大性能損失或復雜度仍較高，并主要集中于對非相關(guān)信道下天線選擇的研究。
針對這一問題，本文以容量性能最大化為目標，提出了基于相異度的天線選擇算法和針對相關(guān)信道下基于CSA的改進選擇算法，并討論了幾種基于智能算法的收發(fā)聯(lián)合天線選擇技術(shù)。主要研究成果如下：
1．介紹了復高斯隨機信道和相關(guān)信道情況下MIMO的系統(tǒng)模型；對MIMO系統(tǒng)容量進行分析，研究信道相關(guān)性對系統(tǒng)容量性能的影響，對幾種經(jīng)典天線選擇算法的容量性能進行比較，仿真結(jié)果表明窮盡搜索算法能夠獲得最優(yōu)的容量性能，但復雜度非常高，經(jīng)典的遞增遞減算法獲得了接近最優(yōu)的容量性能，同時降低了系統(tǒng)復雜度。
2．以實現(xiàn)容量性能最優(yōu)為目標提出了一種基于相異度的天線選擇算法。分析了采用相對誤差、絕對誤差對相異度算法性能的影響，同時通過對幾種算法的復雜度計算分析，證明所提算法保持了較低的運算復雜度。通過仿真將其與窮盡搜索、基于相關(guān)度和相似度等天線選擇算法進行了容量性能比較，仿真結(jié)果驗證了所提算法在容量性能上的改進。
3．針對相關(guān)信道，對基于相關(guān)性的天線選擇算法進行改進，通過選擇具有最小平均相關(guān)性和最大相關(guān)矩陣行列式的天線作為最優(yōu)天線子集。在選擇較少天線的情況，所提算法降低了計算復雜度，達到幾乎與相關(guān)性選擇算法相同的性能，且非常接近最優(yōu)選擇算法。仿真結(jié)果證明了該算法的有效性和可靠性。
4．介紹了基于遺傳算法和模擬退火算法的收發(fā)聯(lián)合天線選擇技術(shù)，并將遺傳算法與模擬退火算法結(jié)合得到一種新的基于智能算法的天線選擇方法。同時，對幾種算法的原理和性能進行了討論，并給出其實驗仿真結(jié)果。
關(guān)鍵詞 多輸入多輸出；天線選擇；相異度；相關(guān)性；遺傳算法；模擬退火
Abstract
Along with the development of wereless communication, the demand of high speed data communication is increasing, conventional single antenna communication system capacity can not meet the requirement of practial application, and the communication system reliability has yet to be further improved. Multiple-Input Multiple-Output (MIMO) technology can utilize multipath effect, and improve the transmission rate and communication quality. It becomes one of key technologies for new generation wireless communication systems. However, MIMO system needs multiple radio frequency (RF) chains for employing multiple antennas, this increase the cost of additional hardware and the complexity of signal processing substantially, and limits development and generalization of MIMO technology to a great extent. Antenna selection technology is to use only an antenna subset of transmit and receive with better performance, it can reduce the expense of hardware and complexity of processing with less performance loss, becomes research focus of MIMO wereless communication field.
As existing antenna selection algorithms in MIMO systems, the exhaustive search algorithm gets the optimum performance, but it requires lots of matrix operations, so leads to very high complexity. Some sub-optimal antenna selection algorithms emerge in order to decrease the complexity of algorithm, but most of these algorithms result in a greater performance loss or still higher complexity, and mainly concentrate on the antenna selection research in the non-correlated channels.
According to the problem, this paper takes the capacity performance maximize for a target, proposes an antenna selection algorithm based on dissimilarity, an improved selection algorithm based on CSA for correlated channel, and discussed several joint transmit/receive antenna selection technology based on smart algorithm. The main results are as follows:
1. MIMO system model in complex Gaussian random channel and correlated channel is introduced, then an analysis of MIMO system capacity and the impact of the channel correlation is given, several classical algorithms capacity performance are compared, the simulation results proved exhaustive search algorithm can obtain the optimal capacity, but the complexity is very high, the decremental algorithm and incremental algorithm get the closed optimal capacity, and lower the system complexity at the meantime.
2. The paper proposes an antenna selection algorithm based on dissimilarity, the algorithm goal is to maximize capacity. It analyzes the impact of relative error and absolute error to the property of dissimilarity algorithm; meanwhile the complexity is computed and analyzed to prove that the proposed algorithm maintains lower computation complexity. The capacity performance of the proposed algorithm, exhaustive search, the algorithms based on ..