Search

We propose a design of ring-core few-mode fiber (RC-FMF) with the erbium doping at the cladding region near the outer ring edge. After the parameter optimization of the RC-FM-EDF, the intensity overlapping difference between LP01 and LP11 modes can be minimized, leading to a DMG of 0.22 dB for the corresponding ring-core few-mode erbium-doped fiber amplifier (RC-FM-EDFA). Meanwhile, in comparison with uniform core doping design, the saturation input signal power of the RC-FM-EDFA can be enhanced from −17.7 dBm and −16 dBm to −8.5 dBm for LP01 and LP11 modes, respectively. Consequently, our proposed RC-FM-EDFA is characterized by both low differential modal gain (DMG) and the enhanced saturation input signal power. Those new characteristics of the RC-FM-EDFA is ideally desired for future agile networks.

With the commercialization of the fifth-generation mobile communication technology, the global network traffic is exploding. Network traffic prediction is studied in order to schedule network resources more efficiently and ensure quality of service. However, the traditional neural network traffic prediction model has poor prediction effect on time series, and cannot accurately predict the high speed and rapid change of network traffic. In this paper, a neural network prediction model is proposed with prediction accuracy improvement of passive optical network traffic, based on Long Short-Term Memory (LSTM) with a new activation function––Swish activation function. According to the characteristics of actual passive optical network traffic, the data simulated by simulation software are tested. Compared with BP neural network and recurrent neural network (RNN), the proposed model is proved to display high prediction accuracy.

A novel approach to interrogate a fiber grating sensor based on spectral-shaping and wavelength-to-time mapping is proposed and simulated. In the system, two chirped fiber Bragg gratings (CFBG), with one used as a reference grating and the other one as sensing grating, are employed to form a Fabry-Perot (FP) interference structure, used as the spectral-shaping filter. The filter can be modeled as a two-tap delay-line filter. An optical pulse from a mode-locked laser source is reflected and spectrally shaped by the FP structure, and then the shaped spectrum is mapped to the temporal domain by a dispersive element. When the sensing CFBG is experiencing a strain, its wavelengths will be changed, leading to different interference spectra. Using Fourier transform, the frequency of the generated microwave waveform can be achieved, from which the wavelength shift of the sensing grating and the applied strain can be demodulated. A mathematical model to describe the microwave waveform generation is developed, by which the interrogation principle is analyzed. Numerical simulations are implemented to verify the proposed approach. The results present that the sensitivity of the system is 0.014 GHz/μe.

A microwave photonic filter (MPF) with reconfigurable frequency responses using a phase-shifted fiber Bragg grating (PS-FBG) of two phase shifts is proposed and a two-passband filter is experimentally demonstrated with tunable frequency spacing. The PS-FBG with two phase shifts acts as an optical notch filter to suppress one sideband of the phase-modulated signals. Thus, the phase-modulated signals are converted to intensity-modulated single-sideband signals. When the wavelength of the optical carrier from a tunable laser source is tuned into different frequency ranges, the frequency responses of the MPF having two narrow passbands, one passband and one notch, or two notches can be achieved. For the dual-passband MPF, contrasting to a single-passband MPF, the advantages in sensor applications are discussed.

Under the opportunity of the government greatly promote Education information in China, guide the high education with modern educational philosophy and support it with modern information technology, Mobile learning would inevitably arouse the education sector's Attention then mobile learning will be an importance complement way to traditional education. However, there are still many difficulties in application mobile learning to high education teaching, only to break through these difficulties can mobile learning integrated into the traditional education better, so that high education more conducive to fostering high-quality personnel.

The purpose of medical image segmentation is to divides the lesions image with a special meaning and background regional. The characteristics of medical images are generally more complex. There is often overlap between different regions, the edge of the area is vague. Research based on fuzzy clustering method for medical image segmentation, fuzzy C-means clustering and fuzzy kernel clustering methods were studied and discussed fully considered the spatial information, the single objective problem conversion into multi-objective problem. The experimental results show that the method use in medical image segmentation achieved good results.

Based on the analysis the advantages and disadvantages of hierarchical clustering and fuzzy clustering methods, make sample points within each subspace as similar as possible, differences between sample points within different sub-space as large as possible, The essence is to find different data models which hidden in different data. It is an unsupervised learning process that enables the blind classification in the sample spaces. Currently clustering data mining research is a very active field.t.

The convergence business has become a new driving force to promote the growth of mobile users. In order to accelerate the development of convergence business, according to the characteristics of the convergence business, this paper designed a new convergence product, analyzed and elaborated the convergence business using Logistic functions and network externality theory, and predicted the development trend. The result shows that the convergence business will enter a stage of rapid development in the future.