空分复用中利用空间光调制器实现对光模式的精确控制和选择方法的研究

发布时间：2018-04-16 12:27

本文选题：空分复用 + 模分复用　；参考：《北京邮电大学》2015年硕士论文

【摘要】：高速光纤通信系统中的空分复用(Space division multiplexing, SDM)是一种全新的光纤信道复用方式。研究空分复用是为了解决日益紧缺的单模光纤的信道容量,它采用的复用方法是在多芯或少模／多模光纤中利用光纤在空间上的自由度来复用信道容量。其中,模分复用(Mode division multiplexing, MDM)是空分复用的一种实现方法,指的是在少模/多模中,采用互相正交的空间模式为信道,从而实现空间复用的方法。模式转换方法是实现模分复用方法的核心技术,实现任意模式间高效、快速的精确转换,对光模式的精确控制和选择方法的研究是模分复用技术进一步推广的基础。近几年,从光纤通信系统中模分复用关键技术研究成果我们可以得出,目前对于模分复用中模式转换方法的设计可以归为两类,即光波导类和空间光路类。模式转换是光纤通信系统中模分复用技术的关键研究点,本论文提出的基于空间光调制器的模式控制与选择方案属于空间光路型模式转换方法,该方法对器件要求低,不需要针对每一种方案特殊定制专一的光路元件,更易于实现。论文首先分析了国内外对于模分复用这一课题的研究进展,分析选题背景和选题依据。然后介绍了课题中涉及到的基本理论知识,包括二维傅里叶变换相关理论、空间频谱、传递函数及透镜的傅里叶变换性质、空间光调制器相关概念等。论文结合了光学傅里叶系统的结构与光场分布模式特征,提出了利用空间光调制器和模拟退火优化技术实现对模式间进行精确控制和选择的方法,并给出了其仿真结果。最后对研究方案的模式转换效果和方案模型的执行效率做横向分析,并提出了局部退火、图形处理器(Graphic Processing Unit, GPU)等进一步优化方案,完善仿真方案模型。结果表明,本文提出的“空分复用中利用空间光调制器实现对光模式的精确控制和选择方法”能够实现任意模式间高效、快速的精确转换。本论文的研究内容和创新点： 1、本论文提出基于光学傅里叶系统和空间光调制器的任意模式间自由转换的仿真方案,并详细记录该方案的MATLAB仿真实现过程。首先利用拉盖尔-高斯函数可以推导出待转换模式和目标模式场分布的理论表达式。在此基础上,基于透镜的傅里叶变换性质能够推导出本仿真平台中待转换模式经透镜1到达空间光调制器左侧的场分布和目标模式经透镜2到达空间光调制器右侧的场分布,方便下一步计算加载到空间光调制器上的模式间传递函数原型。理想的模式间传递函数的表达式是将空间光调制器右侧和左侧的光场分布做除法运算,由此可以知道理想的模式间传递函数是因待转换模式和目标模式的不同而不同的,但是其计算方法都是一样的,至此理论上的模式间自由转换方案能够被验证是可行的。 2、提出基于计算机模拟优化算法的精确模式转换控制和选择方法由于本仿真模型所对应的实验方案选用的是仅能改变加载相位的反射型空间光调制器,而理想的模式间传递函数的表达式是携带幅度和相位信息的复数矩阵,如果简单地去除幅度信息,剩下的相位部分作为模式间传递函数加载到空间光调制器上,会造成信息的大量缺失,模式间转换的精确程度很低。所以如何调整相位信息以实现较为精确的任意模式转换是本论文的一项关键问题。本论文提出了基于计算机模拟优化算法-模拟退火算法对仅有相位信息的模式间传递函数进行随机逐点的相位变更,每一次变更后都依模拟退火算法选择概率接受或放弃变更动作,这样能够对加载到空间光调制器的模式间传递函数实现实时控制,光束通过这种实时变化的模式转换方案模型后,利用模拟退火算法的全局优化性能,能够获得最接近理想目标模式的退火后目标模式,最终实现模式间的精确转换。 3、本论文对精确模式转换方案的实现效果做了多方面分析,通过局部退火、改变终止条件、加入GPU等方式提高转换方案效率和精度。 1)以局部退火代替全局退火：根据理想光斑尺寸,将计算范围缩小为以光斑中心出发的小区域,避免无实际意义区域的运算消耗,传递函数相位图更改范围也做相应修正,可以有效缩短运算时间。 2)退火终止条件自适应化：退火终止条件根据模式和退火效果做自适应化处理,当多次迭代而相关程度没有提升,即使没有达到终止温度也应结束迭代,避免无意义的重复计算。 3)引入GPU：利用专用图像处理核心处理器,实现并行处理。通过上述算法优化处理,算法的精确度和运算效率都有大幅提高,进一步优化了模式精确控制和选择方案。
[Abstract]:Space division multiplexing ( SDM ) in high speed optical fiber communication system is a new mode of optical fiber channel multiplexing .

In recent years , we can conclude that the design of mode conversion method based on spatial light modulator can be classified into two categories , namely optical waveguide type and spatial light path .

This paper first analyzes the research progress of this subject , analyzes the background of selection and the basis for choosing the topic . Then the paper introduces the basic theory knowledge involved in the subject , including two - dimensional Fourier transform correlation theory , spatial frequency spectrum , transfer function , Fourier transform property of lens , spatial light modulator correlation concept , etc . The paper combines the structure of optical Fourier transform and simulated annealing optimization technology to realize precise control and selection between modes . Finally , the paper puts forward some further optimization schemes , such as local annealing , graphic processor ( GPU ) and so on . The results show that the precise control and selection method of optical mode by using spatial light modulator in space division multiplexing is able to realize efficient and fast accurate conversion between arbitrary modes .

The research content and innovation points of this thesis are as follows :

1 . The simulation scheme of free transition between arbitrary modes based on optical Fourier system and spatial light modulator is proposed in this paper , and the realization process of MATLAB simulation of the scheme is recorded in detail .

On the basis of the Fourier transform property of the lens , the field distribution and the target pattern on the left side of the spatial light modulator via the lens 1 can be deduced . The expression of the ideal inter - mode transfer function is to divide the light field distribution on the right side and the left side of the spatial light modulator .

2 . The precise mode switching control and selection method based on computer simulation optimization algorithm are proposed .

In this paper , we propose a computer simulation optimization algorithm - simulated annealing algorithm to carry out random point - by - point phase change of the transfer function between modes . So how to adjust the phase information to achieve more accurate arbitrary mode conversion is a key problem in this paper .

3 . In this paper , the realization effect of the precise mode conversion scheme is analyzed , and the efficiency and accuracy of the conversion scheme are improved by means of local annealing , changing the termination condition , adding the GPU and the like .

1 ) local annealing is used instead of global annealing : according to the ideal spot size , the calculation range is reduced to a small area starting from the center of the light spot , the operation consumption of the non - practical significance area is avoided , the change range of the phase map of the transfer function is also modified accordingly , and the operation time can be effectively shortened .

2 ) self - adaptive annealing termination condition : the annealing termination condition is self - adaptive processing according to the mode and the annealing effect , and when the correlation degree is not improved after multiple iterations , the iteration can be finished even if the termination temperature is not reached , and the meaningless repeat calculation is avoided .

and 3 ) introducing a GPU : utilizing a special image processing core processor to realize parallel processing .

Through the optimization of the algorithm , the accuracy and the operation efficiency of the algorithm are greatly improved , and the mode precise control and selection scheme is further optimized .

【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN929.11

【参考文献】