Virtual draw of microstructured optical fiber based on physics-informed neural networks.

The implementation of microstructured optical fibers (MOFs) with novel micro-structures and perfect performance is challenging due to the complex fabrication processes. Physics-informed neural networks (PINNs) offer what we believe to be a new approach to solving complex partial differential equations within the virtual fabrication model of MOFs. This study, for what appears to be the first time, integrates the complex partial differential equations and boundary conditions describing the fiber drawing process into the loss function of a neural network. To more accurately solve the free boundary of the fiber's inner and outer diameters, we additionally construct a neural network to describe the free boundary conditions. This model not only captures the evolution of the fiber's inner and outer diameters but also provides the velocity distribution and pressure distribution within the molten glass, thus laying the foundation for a quantitative analysis of capillary collapse. Furthermore, results indicate that the trends in the effects of temperature, feed speed, and draw speed on the fiber drawing process align with actual fabrication conditions, validating the feasibility of the model. The methodology proposed in this study offers what we believe to be a novel approach to simulating the fiber drawing process and holds promise for advancing the practical applications of MOFs.

Long short-term memory network of machine learning for compensating temperature error of a fiber optic gyroscope independent of the temperature sensor.

We present an artificial intelligence compensation method for temperature error of a fiber optic gyroscope (FOG). The difference from the existing methods is that the compensation model finally determined by this method only uses the FOG's data to complete the regression prediction of the temperature error and eliminate the dependency on the temperature sensor. In the experimental stage, the proposed method performs temperature experiments with three varying trends of temperature heating, holding, and cooling and obtains sufficient output data sets of the FOG. Taking the output time series of the FOG as the input sample and based on the long short-term memory network of machine learning, the training, validation, and test of the model are completed. From the two perspectives of network learning ability and the improvement degree of the FOG's performance, four indicators, including root mean square error, error cumulative distribution function, FOG bias stability, and Allan variance analysis are selected to evaluate the performance of the compensation model comprehensively. Compared with the existing methods using temperature information for prediction and compensation, the results show that the error compensation method without temperature information proposed can effectively improve the accuracy of the FOG and reduce the complexity of the compensation system. The work can also provide technical references for error compensation of other sensors.

Optical fiber with homogeneous material by side-array cladding.

Optical fibers are the core elements for various fiber-optic applications in communication, lasers, sensors, tweezers, quantum optics, and bio-photonics. Current optical fibers are based on a core-cladding structure with different refractive indices and are mainly fabricated using the stack-draw method. However, such a traditional fabrication method limits the realization of fibers with various advanced optical materials, thereby restricting the utilization of excellent optical properties offered by these materials. In this study, a novel structure for side-array cladding by laser drilling on the side of the fiber with homogeneous material is proposed. Accordingly, the confinement loss, mode characteristics, birefringence, and dispersion of the side-array cladding fiber are investigated based on the numerical simulation performed via the finite element method. Subsequently, an optimal fiber structure is obtained by taking the crystal material as an example. Essentially, our proposed side-array cladding fiber can eliminate the mismatch problem of core-cladding materials in the current stack-draw fabrication method. Potentially, the proposed approach can serve as a standard design and fabrication method of optical fibers with homogeneous material, by utilizing the rapid development of laser processing. In other words, a large number of advanced optical materials can be fabricated into optical fibers with the proposed technique, thus maximizing their technical advantages for different applications.

Pattern recognition using self-reference feature extraction for φ-OTDR.

This paper proposes a pattern recognition method for φ-OTDR based on self-reference features, where machine learning is applied to classify the vibration monitored. The φ-OTDR collects the light amplitude-time-space sequence, establishes a reference position in the spatial dimension, and combines the two dimensions of the vibration and reference positions to form self-reference features, which are then used as machine learning features. These self-reference features can effectively improve the pattern recognition accuracy. This paper selects a low sampling frequency for data collection, analyzes the influence of sample definition methods of different time lengths on the pattern recognition accuracy, and determines that the optimal sample length is 10 data points. The contribution of different feature parameters to pattern recognition is analyzed, and eight eigenvalues such as average, maximum, and minimum are finally determined to form self-reference features that are used as the input of the machine learning algorithm. The recognition accuracies of five machine learning algorithms including kNN, Decision Tree, Random Forest, LightGBM, and CatBoost are analyzed and compared, and the CatBoost algorithm in the integrated learning algorithm is finally determined as the optimal algorithm. On this basis, this paper proposes a filtering algorithm to deal with abnormal signals, which can effectively compensate for abnormal data and further improve the accuracy of pattern recognition. Finally, this paper conducts the pattern recognition study on four common events of tapping, bending, trampling, and blowing, and obtains the average recognition rate of 98%. In addition, this paper innovatively carried out pattern recognition research on five types of mining equipment, including ball mills, vibrating screens, conveyor belts, filters, and industrial pumps, and obtained the average recognition rate of 93.5%.

All-solid anti-resonant single crystal fibers.

In this paper, a novel all-solid anti-resonant single crystal fiber (AR-SCF) with high refractive index tubes cladding is proposed. By producing the cladding tubes with high refractive index material, the AR guiding mechanism can be realized for the SCF, which can reduce the mode number to achieve single-mode or few-mode transmission. The influences of different materials and structures on the confinement loss and effective guided mode number for wavelengths of 2-3 µm are investigated. Then, the optimal AR-SCF structures for different wavelengths are determined. Furthermore, the influences of different fabrication errors are analyzed. This work would provide insight to new opportunities in the novel design of SCFs by AR, which would greatly impact the fields of laser application, supercontinum generation, and SCF sensors.

Discovering extremely low confinement-loss anti-resonant fibers via swarm intelligence.

In this work, we obtain extremely low confinement-loss (CL) anti-resonant fibers (ARFs) via swarm intelligence, specifically the particle swarm optimization (PSO) algorithm. We construct a complex search space of ARFs with two layers of cladding and nested tubes. There are three and four structures of cladding tubes in the first and second layer, respectively. The ARFs are optimized by using the PSO algorithm in terms of both the structures and the parameters. The optimal structure is obtained from a total of 415900 ARFs structures, with the lowest CL being 2.839×10-7 dB/m at a wavelength of 1.55 µm. We observe that the number of ARF structures with CL less than 1×10-6 dB/m in our search space is 370. These structures mainly comprise four designs of ARFs. The results show that the optimal ARF structures realized by the PSO algorithm are different from the ARFs reported in the previous literature. This means that the swarm intelligence accelerates the design and invention of ARFs and also provides new insights regarding the ARF structures. This work provides a fast and effective approach to design ARFs with special requirements. In addition to providing high-performance ARF structures, this work transforms the ARF designs from experience-driven to data-driven.

Use of machine learning to efficiently predict the confinement loss in anti-resonant hollow-core fiber.

The fundamental mode confinement loss (CL) of anti-resonant hollow-core fiber (ARF) is efficiently predicted by a classification task of machine learning. The structure-parameter vector is utilized to define the sample space of ARFs. The CL of labeled samples at 1550 nm is numerically calculated via the finite element method (FEM). The magnitude of CL is obtained by a classification task via a decision tree and k-nearest neighbors algorithms with the training and test sets generated by 290700 and 32300 labeled samples. The test accuracy, confusion matrices, and the receiver operating characteristic curves have shown that our proposed method is effective for predicting the magnitude of CL with a short computation runtime compared to FEM simulation. The feasibility of predicting other performance parameters by the extension of our method, as well as its ability to generalize outside the tested sample space, is also discussed. It is likely that the proposed sample definition and the use of a classification approach can be adopted for design application beyond efficient prediction of ARF CL and inspire artificial intelligence and data-driven-based research of photonic structures.

Association between interleukin 8 -251 A/T and +781 C/T polymorphisms and osteosarcoma risk in Chinese population: a case-control study.

Interleukin-8 (IL-8) is an angiogenic chemokine that plays a potent role in both development and progression of many human malignancies. However, there are no data about the role of IL-8 polymorphism in development of osteosarcoma. A hospital-based case-control study was conducted among 190 patients with osteosarcoma and 190 healthy controls to investigate the possible association between the IL-8 -251 A/T and +781 C/T polymorphisms, respectively, and the risk of osteosarcoma. Significant differences of genotype distribution were observed between osteosarcoma cases and controls at the IL-8 -251T/A genotypes. Compared with the IL-8 -251T/A homozygote TT, the heterozygous TA genotype was associated with significantly increased risk for osteosarcoma (odds ratio (OR) = 2.16, 95 % confidence interval (CI) = (1.38-4.52), P = 0.021); the AA genotype was associated with increased risk for osteosarcoma (OR = 1.94, 95 % CI = 1.31-3.83, P = 0.018). TA and AA combined variants were associated with increased risk for osteosarcoma compared with the TT genotype (OR = 1.72, 95 % CI = 1.45-4.41, P = 0.023). Moreover, the genotype AA of IL-8 -251T/A carried a higher risk of osteosarcoma metastasis and later Enneking stages, compared with the TT genotype. However, the genotype and allele frequencies of IL-8 +781 C/T polymorphisms in osteosarcoma patients were not significantly different from controls. Our results showed that the IL-8 -251 A/T genotype was associated with increased risk for development and metastasis of osteosarcoma in Chinese Han population.