基于SVD和GMD的可见光通信系统的研究

Research on visible light communication system based on SVD and GMD

在多输入多输出(MIMO)的可见光通信(VLC)系统中,双奇异值分解(SVD)的预编码方案得到信噪比(SNR)相差非常大的子信道。针对这一问题,文章提出基于SVD和几何均值分解(GMD)的预编码方案,即先在用户的互补信道矩阵上进行SVD分解,再在等效信道矩阵上进行GMD分解。这个方案既有抑制用户间干扰的优点,又可以得到相同SNR的子信道,并且消除复杂的比特分配过程。文章还推导出方案的理论符号错误率(SER),通过仿真比较得到的SER,验证了基于SVD和GMD的预编码和对称矩阵辅助调制的方案比基于双SVD的预编码方案的数据通信传输质量更高。

In multiple-input and multiple-output(MIMO)visible light communication(VLC)systems,the pre-coding scheme of double singular value decomposition(SVD)obtains subchannels with very large signal-to-noise ratio(SNR)differences.To solve this problem,this paper proposes a precoding scheme based on SVD and geometric mean decomposition(GMD),which first performs SVD decomposition on the user's complementary channel matrix,and then performs GMD decomposition on the equivalent channel matrix.This scheme not only has the advantage of reduc-ing inter-user interference,but also can obtain subchannels with the same SNR and eliminate the complicated bit allo-cation process.The theoretical symbol error rate(SER)of the scheme is derived,and the simulation results show that the scheme based on SVD and GMD and symmetric matrix assisted modulation has higher data communication transmis-sion quality than the scheme based on double SVD-precoding.