基于干扰驱动的多输入单输出认知系统线性预编码设计

Interference-driven Linear Precoding Design in Cognitive Radio Multiple Input Single Output Systems

在多天线(MIMO)认知无线电网络(CR)中,传统的预编码算法主要是抑制或消除认知用户(SUs)之间的干扰,达到优化系统性能的目的。与传统算法不同,该文从利用干扰的角度提出了两种新的预编码算法,即认知部分线性预编码(CR-PLP)和认知相位调整预编码(CR-PALP)。理论分析和仿真结果表明,这两种算法都能够利用星座图中的相位信息,通过保留相长干扰,提高接收端信号的信干噪比(SINR)。与CR-PLP使相消干扰为零不同,CR-PALP通过相位旋转将干扰转化成相长干扰,以较低的计算复杂度,进一步降低了误符号率(SER),提高了认知用户的信息速率。

Conventionally, linear precoding methods in a Cognitive Radio （CR） Multiple Input Multiple Output （MIMO） network mainly aim to limit or completely cancel the interference between Secondary Users （SUs） while achieving superior performance of the CR system. In contrast to the traditional precoding method, this paper presents two novel precoding methods from the perspective of using interference, namely Cognitive Radio Partial Linear Precoding （CR-PLP） and Cognitive Radio Phase Alignment Linear Precoding （CR-PALP）. The theoretical analysis and simulation results show that these two methods both enhance the Signal to Interference and Noise Ratio （SINR） through constructive interference by taking advantage of the phase information of the constellation. Contrary to CR-PLP where the destructive interference is zeroed, CR-PALP rotates and converts it into constructive part. Then less Symbol Error Rate （SER） and enhanced information transmission rate are achieved with lower computational complexity.