傳統用於多載波系統的一維注水演算法僅考慮在頻率上依據不同子載波遭遇通道之特性調整功率配置,然而當多輸入多輸出系統與正交分頻多工系統結合時,傳輸路徑上的差異可能更勝於頻率選擇的效應,因此空間上的多樣性也應做為功率配置的考量,以利通道容量最大化。 在本論文中,我們基於一維之注水原理,於具相關性之通道下針對多輸入多輸出正交分頻多工系統上提出二維注水方法及實際功率配置方法。經由模擬證明在具相關性通道下,二維注水功率配置比一維注水配置法更能有效提升通道容量和降低錯誤率。我們亦證明此二維注水方法適用於具不同傳送和接收天線個數之多輸入輸出正交分頻多工系統中。
Conventional one-dimension waterfilling algorithm adopted on a multicarrier system adjusts the power allocation according to the channel state information in the frequency domain. However, for a MIMO-OFDM system, the divergence caused by different spatial propagation paths may surpass the effect caused by the frequency selective fading. For this reason, we optimize the channel capacity by re-allocating the total power at both the spatial and frequency domain in this thesis. Based on one-dimension waterfilling algorithm, we propose the two-dimension waterfilling method on the MIMO-OFDM system over the correlated channels. We prove that the 2-D waterfilling achieve higher channel capacity and better BER performance than the 1-D waterfilling. Besides, the proposed method is applicable for MIMO-OFDM systems with the arbitrary transmit or receive antennas.