在传统的直接检测光通信系统中,色散和光探测器的平方检测相互作用会带来不可避免的性能损失。基于斯托克斯空间复用的直接检测光通信系统,提出通过频谱压缩算法进行发射信号序列相关的相位优化,将三维斯托克斯信号转化为四维带宽为3/4的琼斯空间信号,以实现无性能损失、无需额外硬件成本的色散预补偿。鉴于上述理想的频谱压缩方法需要复杂度过高的全局优化算法,提出一种基于查表的简化相位优化算法。仿真结果表明,当优化长度为7个符号时,频谱峰值和谷值之间落差约为17 dB,可有效降低发射机带宽需求。

In a traditional direct detection (DD) optical communication system, performance loss is inevitably caused by the interaction between dispersion and square-law detection of a photodetector. Based on a DD optical communication system using Stokes space multiplexing, a spectrum compression algorithm is proposed to optimize the phase of a transmitted signal sequence, which transforms a three-dimensional Stokes space signal to a four-dimensional Jones spatial signal with 3/4 bandwidth, thus enabling the dispersion pre-compensation without performance loss and extra hardware cost. Since this ideal spectrum compression requires global optimization algorithm with high complexity, we propose a simplified look-up-table based phase optimization algorithm. Simulation results show that when the optimization length is 7 symbols, the difference between the peak and the valley of the spectrum is about 17 dB, which effectively decreases the bandwidth requirement of the transmitter.