利用矩阵将数字信号处理算法采用并行化方式表示, 然后利用矩阵张量积的思想, 推导出以张量积表示的两级迭代并行滤波算法。在此基础上进一步分解得到多级迭代的并行滤波算法, 并且采用多相分解的方式进一步提高系统应用中的并行化率。利用上述算法对码率为5 Gbps的16QAM信号进行32路并行化处理的MATLAB乘加级仿真, 仿真结果与串行算法得到的数据等效, 最终得到与理想误码率对比的系统误码率, 仿真结果与理论误码率误差小于0.05 dB。

This paper firstly exhibits digital signal processing algorithm in a parallelization method with matrix representation. A two-stage iterated parallel digital filtering algorithm is expressed based on tensor product. Then a multi-stage iterated parallel algorithm is composed of two-stage algorithm, and a polyphase algorithm is conducted in order to enhance the parallel ratio. A 16QAM signal with 5 Gbps code rate is paralleled to 32 routes with the above algorithm and simulated in MATLAB. The simulation results show that the parallel algorithm is equivalent to the serial algorithm. Finally, the bit error rate is obtained, and the error between the theoretical bit error rate and the simulation results is within 0.05 dB.