Research On Optimizing Silcon-based Waveguide Grating Coupler Based On Chaos Particle Swarm Optimization

The research on silicon-based photonic integrated chips has developed rapidly in recent years and has become one of the most popular research directions in the field of information technology.Grating couplers,as basic devices for coupling light into silicon-based photonic integrated chips,have been widely studied.The performance of grating couplers is the key point of the integrated chips.In this thesis,the research is based on the silicon-based waveguide grating coupler and an optimization design is studied by combining the particle swarm optimization and FDTD numerical simulation technology.In order to solve the problem that the particle swarm algorithm easily falls into local convergence when dealing with multi-parameter complex problems.This paper proposes the chaos particle swarm optimization algorithm based on the chaotic topology and the local optimal solution mutation,which is applied to the optimization of the Silicon-based waveguide grating coupler,achieving good problem solving results.This thesis is introduced in accordance with the following sections:Firstly,the silicon-based waveguide grating coupler is modeled in FDTD numerical simulation software.The characteristic relationship between the grating structure parameters(duty ratio,period,etching depth)and coupling efficiency is studied one by one using the control variable method and the characteristic curve between the two is drawn.After that,particle swarm optimization is used to optimize the parameters of the waveguide grating coupler.The parameter values obtained at the 22%extreme point of the coupling efficiency are consistent with the characteristic curves,verifying the effectiveness of the proposed algorithm in solving such problems.Secondly,the factors that limit the coupling efficiency of the uniform grating are analyzed,and the particle swarm optimization algorithm is applied to the optical field of the non-uniform periodic grating.The coupling efficiency is improved to 45%.When the general particle swarm optimization algorithm deals with complex optimization problems,it will lead to the slower convergence rate in the later period and the optimization accuracy due to the loss of population diversity and the ease of falling into local optimum.To solve this problem,a chaotic particle swarm optimization algorithm based on chaotic topology and local optimal solution mutation is proposed in this paper.And the improved algorithm is used to optimize the non-uniform waveguide grating coupler.The results show that the optimal coupling efficiency reaches 48%under the same algebra and the convergence rate is faster.Finally,the incident light wavelength,duty ratio,period,and etch depth parameters are regarded as the imput,the corresponding coupling efficiency calculated using FDTD numerical simulation is regarded as the output,and BP neural network algorithm is used to establish a 4-9-1 neural network pair to train the sample data.The results show that the correlation coefficient of the training sample is as high as 0.99883 and the mean square error is only 0.089217%.The training accuracy meets the.prediction requirements of the coupling efficiency of the grating coupler on the network,and then the model is tested with the test sample to obtain prediction data,with smaller absolute error of the simulation value.This provides a new idea for device design,which greatly saves working time.In summary,the innovation point of this thesis is mainly reflected on the optimization of non-uniform waveguide grating couplers,analogous to the similarity which the former and multimodal functions finding optimal texts.The particle swarm algorithm is improved based on chaotic topological and local optimal solutions variation which improve the convergence speed and optimization ability of the algorithm.At the same time,BP neural network algorithm is used to predict the coupling efficiency for grating couplers with different wavelengths.The coupling efficiency of unknown structural parameters is obtained based on empirical data,and getting rid of the structure simulation process,reducing the working time.