Font Size: a A A

Research On Interference Cancellation Technology And Performance Optimization Of High-speed Optical Non-orthogonal Discrete Multitone

Posted on:2023-05-18Degree:MasterType:Thesis
Country:ChinaCandidate:C Y XiongFull Text:PDF
GTID:2568307046493444Subject:Electronic and communication engineering
Abstract/Summary:
Non-orthogonal discrete multi-tone modulation(NODMT)compresses the sub-carrier spacing of discrete multi-tone modulation(DMT)signals,improves signal spectral efficiency,and reduces the impact of bandwidth limitations,and has received widespread attention in short-distance optical fiber transmission.However,the compression of the subcarrier spacing makes the subcarriers no longer orthogonal to each other,resulting in the introduction of inter-carrier interference(ICI)between the subcarriers.How to balance performance and complexity and choose an appropriate ICI elimination scheme is still an urgent problem to be solved.If the linear detection with low complexity is used,such as the zero-forcing algorithm and the minimum mean square error algorithm,although the complexity is low,the performance of eliminating ICI is also poor.If the maximum likelihood detection algorithm with high complexity is used,although the performance is optimal,the complexity of the algorithm increases exponentially with the modulation order and the number of sub-carriers.This paper starts with channel coding and continuous interference cancellation algorithm,and proposes and simulates the framework of highspeed optical NODMT signal generation and receiver ICI cancellation based on code-assisted iterative decoding structure.The main works are as follows:(1)A framework for generating high-speed optical NODMT signals and eliminating ICI at receiver is proposed based on code-assisted iterative decoding structure.The time-domain equalizer and the frequency-domain equalizer are applied to the framework respectively.And the time-domain equalizer is designed by an adaptive moment estimation algorithm based on batch gradient descent.The ICI elimination method proposed in this paper has little correlation with the number of subcarriers used,which enables the system to use a large number of subcarriers,so as to achieve better channel balancing effect.The system is combined with the bit and power allocation algorithm to maximize the utilization of channel bandwidth and the signal-to-noise ratio of each subcarriers.In this paper,the Hughes-Hartogs,Chow and Fischer allocation algorithms are respectively applied to NODMT.The allocation scheme and performance are analyzed and discussed,and the Fischer algorithm is modified.The performance of the modified Fischer algorithm is close to Hughes-Hartogs.Simulation results show that the performance of DMT loaded with bit and power is better than that of DMT loaded without bit and power,and the performance of NODMT loaded with bit and power is better than that of DMT loaded with bit and power.(2)Two schemes are proposed to improve the coding rate.The first scheme is to combine the punctured convolution code with Viterbi soft decision decoding,and adjust the framework.Direct decision is made first,and iterative decoding is carried out to calculate ICI and eliminate ICI after the bit error rate drops to the decoding threshold.When the signal bandwidth is compressed to0.85 and 0.9,the coding rate can be increased to 2/3 and 3/4 respectively.The second scheme uses quasi-cyclic low-density parity-check(LDPC)codes for coding.This LDPC can achieve 1/3,1/2,2/3,3/4 and 5/6 coding rates.The receiver uses the LLR belief propagation algorithm,the min sum algorithm and the normalized min sum algorithm respectively for decoding.The simulation results show that when the compression ratio is reduced to 0.925,the coding rate can be increased to 3/4.After the coding rate is increased,the same information is carried by fewer bits,and the redundancy caused by coding is reduced,thereby reducing the overhead.
Keywords/Search Tags:Non-orthogonal Discrete Multi-Tone, Bandwidth limitation, Channel coding, Interference cancellation, Bit Allocation
Related items