Robust beamforming in interference channels with imperfect transmitter channel information
Robust beamforming in interference channels with imperfect transmitter channel information
复制标题
DOI：
10.1016/j.sigpro.2012.03.014
复制
发表时间：
2011-10
期刊：
影响因子：
--
通讯作者：
R. Mochaourab;Eduard Axel Jorswieck
中科院分区：
文献类型：
--
作者：
R. Mochaourab;Eduard Axel Jorswieck
我们认为K个链接同时在相同的频谱带。每个发射器有多个天线，而每个接收器使用单个天线。此设置对应于多输入单输出干扰信道。我们假设在单用户解码接收机完美的信道状态信息，而发射机只有真实信道的估计。信道估计误差被假定为有界的椭圆形区域，其几何形状是已知的发射机。鲁棒波束成形优化最坏情况下的接收功率增益，并且帕累托最优点是最坏情况下可实现的速率元组，从该最坏情况下可实现的速率元组不可能在不降低另一链路的性能的情况下提高链路的性能。我们描述了在任何帕累托最优点操作所需的鲁棒波束形成向量。此外，这些波束形成向量由K（K-1）个实值参数参数化。我们分析了系统在高信噪比和低信噪比下的频谱效率。只有当估计误差与逆SNR成线性比例时，迫零传输才能在高SNR下实现全复用增益。如果误差与SNR无关，则单用户传输在高SNR下是最佳的。在低SNR下，鲁棒的最大比率传输优化了可靠通信的每比特最小能量。数值模拟验证了理论结果。
We consider K links operating concurrently in the same spectral band. Each transmitter has multiple antennas, while each receiver uses a single antenna. This setting corresponds to the multiple-input single-output interference channel. We assume perfect channel state information at the single-user decoding receivers whereas the transmitters only have estimates of the true channels. The channel estimation errors are assumed to be bounded in elliptical regions whose geometry is known at the transmitters. Robust beamforming optimizes worst-case received power gains, and a Pareto optimal point is a worst-case achievable rate tuple from which it is impossible to increase a link's performance without degrading the performance of another. We characterize the robust beamforming vectors necessary to operate at any Pareto optimal point. Moreover, these beamforming vectors are parameterized by K(K−1) real-valued parameters. We analyze the system's spectral efficiency at high and low signal-to-noise ratio (SNR). Zero forcing transmission achieves full multiplexing gain at high SNR only if the estimation errors scale linearly with inverse SNR. If the errors are SNR independent, then single-user transmission is optimal at high SNR. At low SNR, robust maximum ratio transmission optimizes the minimum energy per bit for reliable communication. Numerical simulations illustrate the gained theoretical results.