A Study on transmit power control per subcarrier in OFDM

在多输入多输出系统中，最常见的应用是空分多路复用（SDM），其中各个数据子流从具有相同功率和相同速率的多个天线传输。SDM应用假定MIMO信道信息在发射机处不可用。如果我们知道所有的MIMO信道信息，就可以为每个子流分配匹配的波束，并通过对MIMO信道矩阵的奇异值分解得到波束形成的最佳权值。我们称具有本征波束形成的SDM为E-SDM。在E-SDM中，我们必须考虑资源（传输功率和数据速率）的适应性。在本研究中，我们采用最小误码率（BER）准则作为OFDM系统中每副载波发射功率控制的扩展。结果表明，资源自适应的E-SDM比传统SDM具有更高的误码率性能。然而，在发射机所需的MIMO信道信息包括估计和跟踪误差。在这种误差下，我们还对E-SDM的性能进行了评估。结果表明，在E-SDM和SDM应用中，信道估计误差导致的信道退化几乎相同。当最大多普勒频率非常高时，由传输时间滞后引起的信道跟踪误差会严重降低E-SDM的性能，但我们已经表明，可以通过采用估计信道的一阶外推来补偿它。

In multiple-input multiple-output systems, most common application is space division multiplexing (SDM) where individual data substreams are transmitted from multiple antennas with the same power and the same rate. SDM applications assume that MIMO channel information in not available at the transmitter. If we know all of MIMO channel information, it is possible to assign matched beam to each substream where the optimum weight for beamforming is obtained by singular value decomposition of MIMO channel matrix. We call SDM with eigenbeam-forming as E-SDM, hereafter. In E-SDM, we have to consider resource (transmit power and data rate) adaptation. In this research, we employed minimum brt error rate (BER) criterion as expansion of transmit power control per sucarrier previously used in OFDM systems. As a result, E-SDM with the resource adaptation obtained higher BER performance in comparison to conventional SDM. However, the MIMO channel information required at the transmitter includes estimation and tracking errors. Under such errors, we also evaluated the E-SDM performance. The results show that degradation due to the channel estimation error is almost the same to E-SDM and SDM applications. Although the channel tracking error caused by a time lag until transmission degrades E-SDM performace severely when the maximum Doppler frequency is very high, we have shown that it can be compensated by employing the first-order extrapolation of estimated channel.