Abstract:With the advent of the 5G era, the capacity of existing single-mode optical fiber communication networks is approaching the Shannon limit, making it increasingly difficult to meet the rapidly growing demand for network bandwidth. To address this challenge, space-division multiplexing (SDM) technologybased on expanding the spatial dimension-has emerged as a promising solution. By multiplexing spatial channels within multimode fibers, few-mode fibers, multi-core fibers, or few-mode multi-core fibers, SDM enables an order-of-magnitude increase in transmission capacity per fiber. Simultaneously, it significantly reduces system footprint and deployment costs through efficient spatial resource utilization. Beyond its transformative role in expanding optical communication capacity, SDM has also introduced new paradigms in multidimensional fiber-optic sensing and catalyzed breakthroughs in high-resolution imaging and system flexibility. However, practical implementation of SDM systems is challenged by mode crosstalk, inter-core crosstalk, and mode-dependent loss—mainly arising from fiber and component fabrication imperfections. These factors can lead to severe inter-channel interference and degradation of overall system performance. To ensure that SDM fibers and devices meet stringent performance requirements, the development of high-precision characterization and testing systems is essential. This paper identifies the key parameters for both SDM fibers and associated devices, and provides a comprehensive review of current mainstream testing techniques, including multi-channel optical time domain reflectometry, spatially and spectrally resolved imaging, fixed analyzer methods, coherent optical frequency domain reflectometry, swept-wavelength interferometry, and off-axis digital holography. Each technique exhibits unique advantages in terms of measurable targets, parameter coverage, and application scenarios. This is the first review to systematically compare the strengths, limitations, and suitability of various SDM testing technologies, offering practical guidance for selecting appropriate testing strategies for SDM devices. Finally, the paper presents an outlook on the future development trends of SDM characterization and testing technologies.

Abstract:Micro-electro-mechanical systems (MEMS) gas sensors have emerged as pivotal components for meeting the national strategy of “carbon peak and carbon neutrality,” as well as for achieving distributed, real-time monitoring in fields such as environmental monitoring, industrial safety, and medical diagnostics, due to their core advantages of miniaturization, low power consumption, and mass production. Among the three main categories—chemical, physical, and optical—optical MEMS gas sensors have shown irreplaceable value in specific applications such as detecting flammable, explosive, or highly corrosive gases, as well as trace analysis. Their inherent advantages, such as non-contact, high selectivity, and anti-electromagnetic interference, make them an important development direction for overcoming the limitations of chemical and physical sensors in terms of selectivity, response time, and lifetime. This review systematically summarizes the sensing principles and technical approaches of mainstream MEMS gas sensors, focusing on the core technical bottlenecks faced during the MEMS integration of optical gas sensors. These challenges are specifically reflected in the miniaturization and synergistic integration of three key components: developing efficient and stable miniaturized light sources, miniaturized gas chambers with long effective optical paths, and highly sensitive, low-noise integrated detectors. Based on an extensive literature research, the development and application of new materials, advanced micro-nano manufacturing and structural design, and the fusion of system integration with intelligent algorithms are important directions to break through the current integration challenges of optical MEMS sensors. Furthermore, this study highlights future trends in optical MEMS gas sensors, particularly their progression toward multifunctional integration and intelligent networking, aiming to provide a systematic reference for related research and to help realize a true “lab-on-a-chip” type intelligent sensing system.

Abstract:Flexible assembly has emerged as the dominant technique for the compliant assembly of low-stiffness aircraft wall panels. Critical to this process is the online measurement of key geometric parameters—such as panel positioning targets and real-time surface profiles—to ensure high precision and minimal residual stress. However, the restricted accessibility within the confined, near-enclosed tooling environment, coupled with the inherently poor shape stability of low-stiffness panels, renders direct measurement methods inefficient and prone to errors, creating significant obstacles for real-time monitoring. To address these challenges, this paper introduces a computational measurement approach for estimating key geometric parameters during flexible panel assembly. By combining sparse measurable data with physics-based analytical models, the method enables real-time estimation of parameters that are otherwise difficult to access directly. Specifically, a modal superposition-based technique is developed to compute the global displacement of positioning tooling from limited displacement measurements, facilitating continuous tracking of panel positioning targets. Additionally, a surface profile reconstruction method is proposed, which integrates geometric constraints and the displacement-strain relationship to fuse discrete strain data for real-time shape estimation. Validation was conducted on a downscaled platform simulating the flexible assembly of aircraft horizontal tail panels. The results demonstrate that the proposed method achieves positioning target measurement errors below 75.1 μm and surface profile errors under 18.11%, with computation times shorter than 0.005 s and 0.01 s, respectively. This method provides reliable real-time data support for compliant aircraft panel assembly and advances the development of flexible assembly technologies in aerospace manufacturing.

Abstract:Photoacoustic technology, as a non-destructive detection method that integrates the advantages of both optical methods and acoustic techniques, has been widely applied in various fields, including spectral analysis, imaging diagnostics, and material testing. Benefiting from its broad spectrum and high resolution, the dual-comb has attracted increasing attention in the field of photoacoustic measurements in recent years. However, current researches are primarily focused on absorption spectroscopy, especially for gas detection, while studies on solid materials remain limited. To overcome this limitation, a PA measurement method using the dual-comb light source as the excitation is introduced. Based on two SESAM-based mode-locked Yb3+ doped fiber lasers with tunable repetition rates as seeds, 775 nm high-coherence dual optical frequency combs are generated via nonlinear frequency conversion, including OPA and SHG. Using high-purity (99.99%) tungsten as the target sample, the photoacoustic signal excited by the dual-comb is recorded to measure tungsten′s absorption spectrum at 775 nm, and the corresponding relationships between the excitation signal and the PA signal are validated in both the time and frequency domains. Furthermore, the system is applied to measure the morphological features of tungsten spheres (diameter 168.60 μm) and tungsten wires (diameter 1 143.03 μm) using a three-dimensional translation stage. The relative errors between the measured values and the actual dimensions are calculated to be 11.44% and 3.552%, respectively. For the first time, the photoacoustic spectroscopy and morphological measurements of solid samples are realized by dual-comb and PA effect, confirming the feasibility of the dual-comb for solid morphology measurements and providing a novel technical approach for non-destructive testing of solid materials and devices.

Abstract:The rotating laser-scanning system achieves spatial three-dimensional coordinate measurement based on the principle of angular intersection, which has been widely applied in the fields of aerospace, high-end shipbuilding, and other large-scale equipment manufacturing due to its capabilities for multitarget parallel measurement and scalable measuring range. However, the limited installation space for lasers in existing instruments constrains the selection and optimization of the system′s laser. This article devises the main optical structure of the novel transmitter by analyzing the optical path design requirements. The laser optical path is extended to the exterior of the transmitter via optical fiber conduction. A mirror set is utilized to split light and generate scanning light planes. In combination with a turntable, a precise spatial light field is constructed, which circumvents the limitation of installation space on laser performance, enabling unconstrained subsequent optimization design of the light source and effectively reducing the impact on system performance at the optical signal level. Due to the unique structure of the newly designed laser-guiding transmitter, structural errors can lead to instantaneous changes in the scanning light plane shape during the operation of the transmitter. As a result, traditional parameter models based on rotating light plane characteristics are no longer applicable. Therefore, this article integrates the geometric structure of the optical path, analyzes errors such as optical axis eccentricity and optical axis tilt, abandons the definition in the traditional model that takes the initial spatial equations of the light planes as internal parameters, redefines the internal parameters of the transmitter, and constructs a multi-station networking positioning model and an internal and external parameters calibration model suitable for the laser-guiding transmitter. Finally, an experimental platform is established for simulation and experimentation. The results show that the measurement model studied in this article is effective and has a coordinate measurement accuracy better than 0.7 mm, which meets the requirements of large-scale equipment manufacturing and laying a solid foundation for performance improvement and subsequent optimization of the novel rotating laser-scanning system.

Abstract:Thermal deformation and aerodynamic loads cause axial displacement of shrouded turbine blades, resulting in capacitive sensors deviating from their calibrated positions and thus substantially diminishing the accuracy of labyrinth tip clearance measurements. To tackle this challenge, this study introduces a synchronous measurement method for blade tip clearance and axial displacement using a V-shaped capacitive probe. Initially, theoretical analysis highlights the inherent limitation of traditional circular-core capacitive sensors in detecting the direction of axial displacement due to their symmetrical design, and a nonlinear model is developed to quantify the measurement error induced by axial displacement on blade tip clearance. Subsequently, a V-shaped core capacitive probe is proposed to decouple axial displacement effects from tip clearance measurements by leveraging characteristic waveform parameters such as peak-to-peak voltage (Vpp) and waveform area (Scw). A binary polynomial mapping model is established to associate these parameters with blade tip clearance and axial displacement values. By integrating adaptive filtering and a third-order sinusoidal fitting algorithm, accurate extraction of both parameters is achieved. A dynamic experimental platform is built, and two-dimensional calibration and validation are performed within axial displacement ranges of ±1 mm and tip clearance from 0.5 to 1.5 mm. The results demonstrate that the blade tip clearance measurement achieves an accuracy better than 8 μm—an improvement of 96.8% compared to traditional capacitive methods—while axial displacement is measured with an accuracy of 21.6 μm, including directional identification. Comparative tests with conventional circular-core probes reveal a reduction in tip clearance measurement error under axial displacement from 0.25 mm to 8 μm, validating the effectiveness of the proposed approach. This method offers a robust solution for real-time monitoring of shrouded blade tip clearance under axial displacement, providing significant engineering benefits for intelligent engine operation and active clearance control development.

Abstract:To explore the key parameters of spray flow and fuel combustion characteristics under critical or supercritical pressure conditions in advanced engines, a novel TSCST experimental setup was developed. The main components include a mechanical structure and an integrated gas flow system, featuring several innovative designs: First, a shock dumper with multiple diffusion channels replaces the conventional single-pulse shock tube dump tank, effectively regulating secondary pressure rises caused by additional shock pulses. Second, leveraging shock wave dynamics theory, a three-dimensional converging structure with a specialized curved contour was constructed, substantially enhancing the intensity of the reflected shock wave. Third, an aerosol inlet end cap structure was designed to precisely control the injection of gas-phase fuels as well as low-saturation vapor pressure, high-boiling point, or multi-component complex liquid fuels. Finally, to reduce wall boundary layer effects, the shock tube features a large diameter of 210 mm for both driver and driven sections, making it one of the largest external chemical shock tubes nationwide. Extensive repeatability tests under single- and double-diaphragm conditions have validated the shock wave generation and enhancement capabilities of this setup. Results show that the large-diameter design produces near-ideal pressure profiles lasting up to 14.5 ms. The converging contraction section increases the Mach number of planar shock waves by a factor of 1.6. Additionally, the shock dumpers reduce the pulse pressure of secondary reflected shock waves by over 20% compared to traditional dump tank designs. Overall, this innovative shock tube system exhibits excellent performance in generating high-intensity shock waves while suppressing additional shock pulse effects, offering a powerful experimental platform for investigating high-pressure spray flow dynamics and fundamental combustion reaction kinetics.

Abstract:Underwater acoustic wave vector direction estimation is a crucial task in hydro acoustics, holding significant importance for underwater target identification, detection, and tracking. However, conventional detection methods exhibit inherent limitations. For hydrophone arrays, they are prone to phase ambiguity and acoustic interference. Vector hydrophones have constrained angular resolution and bandwidth due to directivity patterns and resonance frequency limitations. To overcome these physical constraints in traditional detectors, a laser-based acoustic sensing method is proposed. When traversing an acoustic-modulated medium, laser beams undergo measurable deflection without perturbing the original acoustic field. This enables non-invasive acoustic sensing through optical means, circumventing the resonance frequency limitations inherent in conventional transducers. Through rigorous analysis of acousto-optic deflection principles, the mechanism for acoustic vector direction perception is systematically investigated. Theoretical analysis demonstrates that a single-layer laser array configuration can eliminate phase ambiguity during acoustic detection. Building on the non-invasive nature of laser-based sensing, an innovative dual-layer laser array architecture is developed. This enhanced model employs joint correction algorithms to mitigate the influence of sound velocity parameters on wave vector estimation while achieving significant resolution improvement. Experimental results show that both configurations exhibit high angular resolution. Notably, compared with the single-layer array, the dual-layer configuration improves wave vector direction estimation resolution from 0.19° to 0.13°, representing a 31.6% performance enhancement.

Abstract:In jungle environments characterized by complex terrain and dynamic conditions, traditional radio communications are prone to interference, interception, and detection. Bio-inspired acoustic covert communication has emerged as a promising stealth alternative; however, challenges such as severe multipath fading and insufficient concealment remain when mimicking animal vocalizations. To address these limitations, this study proposes a bionic camouflage covert communication method that integrates differential time-delay coding, chaotic encoding-based frame-length modulation, and Multi-Element Virtual Time Reversal Mirror (VTRM) channel equalization. The proposed method constructs a camouflage communication frame comprising synchronization and information codes. Synchronization is achieved using white-eye bird calls in combination with a chaotic encoding mechanism that dynamically modulates frame length, enhancing signal camouflage and unpredictability. Cricket chirps are utilized as information codes to implement differential time-delay coding and correlation-based timing decoding. To counteract multipath-induced signal distortions prevalent in jungle environments, VTRM technology is employed for effective channel equalization. A prototype system was developed and subjected to field testing in realistic jungle conditions, including evaluations of concealment performance. Experimental results demonstrate that under high-concealment scenarios, the proposed method achieves a bit error rate below 1.7×10-3 within a 90-meters range and maintains an average communication rate of 54.2 bit/s. These findings validate the method′s ability to balance concealment and communication efficiency in complex jungle acoustic environments.

Abstract:The noise equivalent index of a magnetohydrodynamics (MHD) angular vibration sensor is a comprehensive evaluation of the scale factor and output noise in the bandwidth range. Specifically, lower output noise and a lower low-frequency cut-off frequency within the -3 dB range indicate better practical performance of the sensor. Due to the extremely low source impedance of the MHD angular vibration sensor, noise matching can be achieved by using transformer coupling, improving the signal-to-noise ratio and significantly reducing the output noise. In this paper, the optimal turn ratio is analyzed and the transformer complete noise output model is established. The introduction of a transformer can increase the low-frequency cutoff frequency of the sensor. At the same time, in practical applications, it is observed that the sensor′s internal leakage magnetic field affects the transformer core, substantially reducing its magnetization inductance. This, in turn, degrades the low-frequency performance and prevents the noise equivalent index from reaching the expected level. To address this issue, a segmented and arc-shaped magnetic shielding structure is proposed for the noise-matching transformer, tailored to the sensor′s unique structure that converts axial magnetic fields into radial magnetic fields. Within the limited spatial constraints of miniaturized sensors, this design effectively mitigates the impact of magnetic leakage on the transformer and enhances the uniformity of the magnetic field within the fluid region. The effect is verified through finite element simulations and experimental testing. This structure greatly improves the low-frequency performance of the sensor in practical applications and significantly optimizes the noise equivalent index. The experimental results show that the low cut-off frequency is reduced to below 2 Hz, the noise equivalent angular rate reaches 2.6 μrad/s RMS, and the noise equivalent angular position is reduced to 32 nrad RMS.

Abstract:Owing to the unique material properties and complex geometry of on-orbit formed space truss structures, traditional 3D visual inspection methods often fail to achieve complete model reconstruction, thereby compromising defect detection accuracy. This paper presents a visionbased detection method for truss structures that leverages 3D curve network graph optimization and designs a rotating visual scanning system for 3D inspection. This system enables comprehensive structural reconstruction and precise defect localization, addressing challenges in on-orbit manufacturing quality control. First, image curve feature recognition is combined with a curve feature matching approach that incorporates distance criteria and curve consistency constraints to establish correspondence between weak-textured slender objects across sequential image frames. Second, by leveraging the principle of 3D curve network graph structure optimization, the method iteratively updates the camera pose and the target structure, thereby computing and refining the geometric topology of the truss while estimating member diameters. Furthermore, a defect detection method based on 3D curvature analysis is introduced and complemented with 2D image validation, enabling accurate identification of typical defects such as virtual joints and breakpoints. Experimental results indicate that, compared to traditional point-feature-matching-based 3D reconstruction methods, the proposed approach achieves superior reconstruction quality and efficiency for slender truss structures. Both geometric parameters and defect detection accuracy exceed 0.3 mm, meeting the requirements for on-orbit manufacturing site inspections. Notably, this method does not rely on external calibration or background features, providing a robust technological foundation for quality control in the on-orbit manufacturing of complex space structures.

Abstract:As a pivotal technology to ensure the safety of oil and gas wells during the entire production cycle, to enhance the efficiency of resource development and to support the strategy of carbon capture and sequestration, the research of cased well cementing quality evaluation closely matches the needs of the national major strategies. It remains a critical and challenging issue in the field of oil and gas engineering. Sonic logging technology has been extensively adopted due to its high sensitivity to the acoustic impedance characteristics of the medium, the integrity of the cement annulus, and the state of the interface. However, the detection sensitivity of traditional acoustic logging technique is mainly focused on the casing-cement interface, and the classical elastic wave theory only describes the fluctuation phenomenon under ideal conditions, limiting its ability to simulate the weak cemented interface state. To address the limitations, the present study focuses on single cased wells as the target, aim to characterize the weak cementation features of different interfaces as slip boundary conditions. The acoustic wave logging model of slip interface theory is constructed, and the propagation characteristics of multimodal guided waves in different interfacial cementation states are investigated based on the array sonic log. A new method is proposed for characterizing the damage of the cement annulus by using the dispersion features of guided waves. Experiments on single cased wells with different interfacial cementation states are carried out, and the data results effectively verified the possibility of evaluating cementing quality through dispersion characteristics. The findings demonstrate that the cementing damage and mechanical properties can be inverted and evaluated according to different guided waves dispersion characteristics, offering valuable insights into wave propagation under complex cementing damage conditions. Furthermore, this provides a theoretical basis for comprehensively utilizing multimodal guided wave propagation characteristics in the cased wells, contributing to both cementing quality assessment and evaluation of cement seal effectiveness.

Abstract:Coordinate measuring machines (CMMs), as core equipment in precision measurement and advanced manufacturing, have their measurement accuracy constrained by both geometric and non-geometric errors. Traditional error compensation methods based on laser interferometry suffer from low efficiency and incomplete compensation models, failing to meet the increasingly stringent accuracy requirements. To address these challenges, this paper proposes a comprehensive error compensation approach that integrates geometric error modeling with neural network techniques to achieve efficient and precise correction of complex errors. For geometric errors, an error model is constructed based on rigid body kinematics, and an adaptive differential evolution algorithm is employed to realize high-precision parameter identification. For non-geometric errors, a novel neural network leveraging neighborhood error features is designed, which utilizes a multi-head self-attention mechanism to deeply capture the error distribution characteristics within the measurement space. Compared to traditional networks that rely solely on the target point position as input, the proposed network significantly improves prediction accuracy. Experimental results demonstrate that, after compensation using the proposed method, the maximum probing error of a CMM with a nominal accuracy of 2.8+L/400 μm is reduced to 0.35 μm, and the length measurement error is improved to 0.5+L/400 μm, showing clear enhancement over the factory-specified accuracy and fully validating the method′s applicability and practicality. Furthermore, compared with conventional compensation techniques, the proposed approach exhibits significant advantages in compensating both geometric and non-geometric errors, achieving robust, efficient, and accurate comprehensive error compensation. This method offers a feasible and effective solution for error control and performance optimization in CMMs and other precision measurement instruments, holding substantial engineering significance and broad application prospects.

Abstract:As a key load-bearing and motion hub of the human body, the knee joint has features of high load capacity and strong stability. Structurally, it consists of the femur, ligaments, and tibia, with multiple ligaments connecting the femur and tibia to share vertical loads and convert part of the compressive stress on the tibia into tensile stress along the ligament direction, thereby preventing overload-induced buckling. Inspired by this biomechanical principle, this study proposes a multi-link bionic inverted pendulum, which consists of four symmetrically distributed C-shaped flexible arms and a rigid platform. This design disperses loads, mitigates stress concentration, and converts compressive stress on the pivot into tensile stress, significantly enhancing stability margins. A stability model is formulated, and the effects of disturbances, such as ground vibrations and centroid offset, are analyzed via simulation. Experimental results show that, under an 8 kg load, the system achieves a resolution better than 0.6 μN, a measurement range of 0.6~1 210 μN, and background noise below 1.42 μN/Hz1/2 in the 0.1 mHz~5 Hz band. Thrust measurements with a micro-Hall thruster show accurate response and a clear linear correlation among thrust, propellant flow, and discharge voltage. This knee-inspired bionic design offers a novel approach for developing high-precision micro-force measurement devices under heavy-load conditions.

Abstract:Aiming at the problem of cumbersome process and low accuracy in the calibration of single-axis turntable, based on the in-depth analysis of the principle of turntable calibration, this paper presents a new type of coded stereo target and a multi-constraint optimization method for high-precision turntable parameters. Firstly, the target integrates three square planes with different positions and poses of coded feature points. The uniqueness of the code can accurately determine the pose changes of each plane. The parameters of the turntable can be calibrated with only a single adjustment of the target pose, significantly simplifies the calibration process. Secondly, an encoding feature point recognition and processing algorithm is designed. By generating an adaptive elliptical mask, the algorithm robustly separates and accurately identifies the coded feature points for reliable decoding, demonstrating strong robustness in various conditions. Finally, a multi-constraint optimization method for high-precision turntable parameters is designed. The objective function is established by combining point constraint, coplanar constraint and normal vector angle constraint, with appropriate weighting parameters to improve calibration accuracy. The experimental results show that the proposed method achieves a calibration accuracy of 0.021 mm with only a single adjustment of the target pose—an improvement of approximately 36.4% compared to traditional methods. Additionally, when assembling a standard sphere based on the calibration results, an accuracy of 0.025 mm is achieved, reducing the error by about 16.7% compared to conventional algorithms. This further validates the effectiveness of the proposed method. Moreover, under varying levels of noise, the optimization algorithm consistently produces stable results, demonstrating excellent robustness.

Abstract:The cesium atomic fountain clock is a standard frequency signal generation device based on the internal quantum transition of atoms and is currently the universal frequency standard. The ultra-low temperature cesium atomic fountain clock is a special type in which cesium atoms operate at liquid nitrogen temperature(80 K), which effectively reduces the blackbody radiation frequency shift to less than 1×10-16, nearly eliminating the influence of this shift. The cavity phase frequency shift and background gas collision frequency shift of the cesium atomic fountain clock will be greatly improved, with the potential to reduce the systematic frequency uncertainty of the cesium atomic fountain clock to below 2 × 10-16. The magnetic shielding device is an important part of the physical system of the fountain clock. The performance of the magnetic shielding device affects the magnetic field frequency shift and uncertainty index of the fountain clock. Since a liquid nitrogen Dewar interlayer is set on the periphery of the resonant cavity and the atomic flight area, the liquid nitrogen inlet and outlet pipes need to penetrate multiple layers of magnetic shielding. The magnetic shielding structure of the ultra-low temperature cesium atomic fountain clock is more complex than that of the room temperature fountain clock, and the development difficulty is greater. In this paper, according to the spatial structure and application requirements of the ultra-low temperature cesium fountain clock, a corresponding magnetic shielding device is designed. The optimal parameters such as the number of layers, thickness, size, and interlayer spacing of the magnetic shielding device are determined by simulation calculation. The magnetic shielding device is processed according to the design. The processed magnetic shielding device is subjected to high-temperature demagnetization and AC demagnetization treatment. After testing, the radial magnetic field change in the core area of the magnetic shielding device is less than 0.1 nT, the axial magnetic field change is less than 1.3 nT, the shielding factor at both ends is greater than 10 000, and the central shielding factor is greater than 60 000, which meets the application requirements of the ultra-low temperature cesium atomic fountain clock.

Abstract:This study proposes an innovative wide-field time-domain diffuse optical tomography (TD-DOT) approachthat integrates spatial frequency domain (SFD) imaging, single-pixel imaging, and time-correlated single-photon counting techniques to enable wide-field time-resolved measurements in turbid media. Furthermore, a TD-SFD-DOT reconstruction method is developed to achieve three-dimensional reconstruction of the target′s optical parameters. The imaging system is built upon a novel single-pixel SFD architecture, employing two spatial light modulators to respectively realize wide-field structured illumination and wide-field detection via a single-photon avalanche diode detectors. The time-correlated single-photon counting technique is embedded into this architecture to achieve high-cost-effective time-resolved wide-field data acquisition. On the algorithmic side, a time-domain SFD diffusion model incorporating internal source terms is proposed to accurately describe the radiative transfer of spatial modulated light in turbid media. This significantly improves the simulation accuracy of photon propagation in near-surface regions and early-time regimes, addressing the limitations of traditional diffusion equations in modeling highly scattering media near the boundary. To fully harness the rich information embedded in time-resolved signals and improve computational efficiency, a TD-SFD-DOT reconstruction algorithm based on overlapping time-gated data types is developed. This enables the decoupled and depth-resolved reconstruction of both the absorption coefficients and the reduced scattering coefficient. Finally, the proposed system and reconstruction method are validated through a series of agar phantom experiments. Results demonstrate that the method achieves accurate three-dimensional quantitative reconstruction of absorption and reduced scattering coefficients at depths of 3~4 mm beneath the surface of turbid media, providing a novel technical solution for tissue optical tomography.

Abstract:The quality of resistance spot welding directly affects the structural stability and safety of automobile bodies. Pixel-level segmentation maps of welding defects are crucial for accurately analyzing defect morphology and severity. To address the limitations of traditional object detection methods in precisely segmenting small-scale defects and achieving high classification accuracy, this paper proposes a precise localization and segmentation method for RSW defects based on a multi-scale feature fusion network. By integrating cross-level feature connections and multi-scale feature matching, the network captures both global welding characteristics and fine-grained defect details, enabling accurate semantic segmentation of defects in large scenes and improving classification accuracy in RSW regions. A candidate region generation network is designed to fuse low-level detailed features with high-level semantic information, and a custom localization loss function is introduced to ensure accurate positioning of spot weld regions. Subsequently, a defect segmentation and localization network is proposed, which incorporates ROI Align and multi-scale feature matching to construct a normal feature bank for spot welds and formulates an anomaly scoring function for pixel-level anomaly scoring of weld regions. Experimental results show that, compared with traditional object detection models, the proposed method improves the classification accuracy for small RSW targets by 25.35% and enhances the F1 score by 14.81%. Moreover, it produces high-precision pixel-level segmentation maps, achieving a Pixel AUROC of 0.94, demonstrating excellent defect recognition capabilities. The method also achieves good performance on open-source RSW datasets from various industrial scenarios, with an F1 score of 0.93, verifying the generalization ability of the model.

Abstract:Circular grating is an important tool to achieve ultra-high precision full circle Angle measurement, which plays an important role in micro and nano machining, aiming and positioning, metrology science and other fields. The circular grating self-calibration technique overcomes the problems of small measuring range and lack of higher reference datums by arranging reading heads and using specific data processing methods to separate systematic errors, thereby enabling self-calibration. To improve the Angle measurement accuracy of circular gratings and reduce the harmonic suppression in circular grating self-calibration, a multi-reading-head self-calibration method is proposed to increase the harmonic residual order. Based on the transfer function method, the relationship between the residual order and the Angle of the reading head is studied, and the formula for calculating the Angle of the multi-reading head with the residual order of 360 integers is established. The calculation results show that for the data sampled from 360 points, the residual of layout 1 in the comparison experiment is 0.8″, and the residual order is an integer multiple of 6. Layout 2 has a residual of 0.2″, and the residual order is an integer multiple of 40. The proposed layout 3 obtained by this method has a residual of 3×10-13″, and the residual order is an integer multiple of 360. By conducting a self-calibration experiment with six reading heads, the proposed method eliminates 94.37% of the errors in the circular grating angle measurement system, achieving full-circle calibration with excellent repeatability. This demonstrates the method′s effectiveness in practical applications. This calibration method significantly enhances the order of harmonic residual in the self-calibration of circular gratings, which offers valuable reference for mitigating harmonic errors in angular measurements and supports the development of ultra-high precision angular displacement sensors.

Abstract:With the continuous advancement of underwater full-duplex communication technology, in the co-frequency simultaneous transmit-and-receive mode, high-power transmitter signals can easily couple into the receiver through multiple paths, generating strong self-interference (SI) that overwhelms the desired signal and significantly degrades communication quality. Particularly in complex underwater acoustic channels, this SI phenomenon becomes more severe, posing a critical bottleneck for improving the performance of underwater communication systems. Therefore, effective suppression of co-located self-interference has become a crucial technical challenge that must be addressed in underwater transceiver systems. To overcome the limitations of traditional digital-domain SI cancellation methods—where the least mean squares (LMS) algorithm suffers from insufficient estimation accuracy and the recursive least squares (RLS) algorithm exhibits high computational complexity—this paper proposes a digital-domain self-interference suppression method based on the stable fast transversal recursive least squares (SFTRLS) algorithm. By introducing forward and backward prediction structures, the proposed algorithm transforms the complex matrix operations in conventional RLS into one-dimensional vector operations, effectively reducing computational complexity from O(N2) to O(N). This approach significantly lowers computational overhead while maintaining excellent SI suppression performance. Extensive simulations under varying signal-to-noise ratios (SNRs) and channel orders demonstrate the proposed algorithm′s advantages in convergence speed, SI channel estimation accuracy, and computational efficiency. Furthermore, real-world lake experiments validate the method′s engineering feasibility. The results show that, under the given test conditions, the algorithm achieves an SI suppression ratio of up to 30 dB. Comparative evaluations with existing methods further confirm that the SFTRLS algorithm exhibits strong applicability and practical value in complex underwater environments. This research not only addresses a key technical challenge in underwater communications but also provides a valuable reference for self-interference suppression in other domains.

Abstract:In stepped eddy current thermography nondestructive testing, the structure of the electromagnetic exciter has an important influence on the distribution and strength of the electromagnetic field and heat transfer field. In this article, an integrated optimization system is designed and applied to the square electromagnetic exciter structure to improve the comprehensive detection performance of volumetric defects in oil and gas storage tanks or large-diameter pipe sections. First, the influence of the variation of each structural parameter of the exciter on the detection performance is analyzed, and the structural parameters with significant cumulative influence are selected as optimization variables by principal component analysis. The objective assignment method is used to give weights to nine detection performance indicators according to the amount of information, and three comprehensive indicators about temperature rise, uniformity and detection efficiency are obtained. Then the discrete optimization space constructed from the key structural parameters about the composite indexes is mapped to the continuous optimization space of the machine learning model. A multi-objective particle swarm algorithm is utilized to obtain the Pareto front. Finally, the optimal structures are ranked according to the approximate ideal solution ranking method. Compared with the initial structure, the optimized structure improves the temperature rise, uniformity, and efficiency indexes by 18.88%, 2.46%, and 73.61%, respectively. compared with the initial structure. In addition, a set of experimental systems is established to verify that the optimized structure has significantly better detection performance than the comparative exciter and can be used to detect defects in steel plates with a thickness of 8 mm and a diameter-to-thickness ratio greater than 3. By integrating and optimizing the electromagnetic exciter structure of the system design, it significantly improves the stepped eddy current thermography detection performance, and experimentally verifies its efficient detection capability of internal volume defects in tank materials, providing an effective solution for the detection of volumetric defects in oil and gas storage tanks and large-caliber pipeline segments.

Abstract:The accurate measurement of dynamic liquid depth is a key technology to analyze the production and operation dynamics of oilfields, and to formulate and adjust oilfield development plans. Because of the conventional pontoon method and acoustic wave reflection method, which cannot be applied to inclined wells due to the influence of the casing annular space, the dynamometer calculation method has poor stability due to the influence of continuous alternating load, and a large amount of calculation is required in the dynamic liquid level prediction method. Starting from the dual-source characteristics of electric energy as both a power source and an information source, a soft measurement method is proposed to correlate the cross-domain correlation between the working electrical parameters of the above-ground motor, the suspension load of the pumping unit, and the dynamic surface depth of the underground oil well. Firstly, according to the energy flow and transfer mechanism of the beam pumping unit during the pumping process, a mathematical model of the donkey head suspension load is formulated by the input electrical parameters of the ground motor, the parameters of the four-link mechanism, and the inclination displacement of the beam. Secondly, based on the fact that the pressure drop of the oil flow in the fixed valve of the pump presents zero value at the moment of opening of the fixed valve of the upstroke and the process of the downstroke, the area between the peak point of the first wave and the trough point of the first wave is dynamically intercepted as the optimal. Finally, according to the correlation between the depth of the dynamic liquid level and the static load on and below the suspension point of the pumping unit, the mathematical model of soft measurement of dynamic liquid level is established, and the input electrical parameters of the collected motor are substituted into the calculation to obtain the soft measurement value of dynamic liquid level depth. The results of engineering experiments and applications show that the method has good stability and strong engineering practicability. It can be applied to electric drive wells. The relative measurement error is not more than ±8%.

Abstract:As a key transmission component in mechanical systems, gearboxes often work under harsh environmental and heavy-load conditions, making them prone to damage. In recent years, built-in encoder signals have emerged as an effective method for health monitoring of mechanical equipment due to their low cost and ease of acquisition. However, the fault-related features within encoder signals are typically weak and often masked by various interference components, which poses a significant challenge for accurate identification and extraction of fault-induced impulses. To address this issue, this article proposes a fault feature extraction and diagnosis method for encoder signals based on a weighted bi-domain sparse decomposition model. In the proposed model, the morphological differences between fault impulses and interference components are analyzed in both the time and frequency domains. By introducing weighted coefficients, periodic binary vectors, and non-convex penalty functions, the model constructs two dedicated non-convex regularization terms that enforce periodic group sparsity in the time domain for fault impulses and spectral sparsity in the frequency domain for harmonic interferences, thereby achieving accurate separation and representation of target features. Furthermore, an iterative solving algorithm for the weighted bi-domain sparse decomposition model is developed by integrating the alternating direction method of multipliers and the majorization-minimization strategy to enhance computational efficiency and convergence stability. Finally, the effectiveness of the proposed weighted bi-domain sparse decomposition model is validated using both simulated data and experimental data from a planetary gearbox test bench. The results indicate that, compared with two existing methods, namely the tunable Q wavelet transform-based decomposition method and the maximum correlated kurtosis deconvolution method, the proposed method shows a significant advantage in fault impulse feature extraction accuracy, demonstrating strong potential for practical engineering applications.

Abstract:Under variable-speed conditions, bearing fault signals exhibit significant non-stationarity and low signal-to-noise ratio (SNR). Traditional sparse representation methods based on static dictionaries are typically designed for fixed-speed scenarios, making them poorly suited to adapt to speed-induced signal variations, which leads to a notable decline in fault feature extraction accuracy. To address this problem, this paper proposes a R-nyi entropy-driven adaptive structural dictionary learning algorithm (RE-ASDLA), aiming to enhance the adaptability and diagnostic accuracy of dictionary learning under non-stationary conditions. The method constructs an overcomplete structural dictionary with time-varying responsiveness, tailored to the transient impact patterns of bearing faults under variable speeds. It overcomes the limitations of conventional segment-based analysis and enables precise extraction of sparse transient features from strong noise backgrounds. During the dictionary update process, a joint optimization of reconstruction error and R-nyi entropy is performed to adaptively refine dictionary parameters, enhancing sensitivity to time-varying features and establishing a time-frequency-sparsity collaborative diagnostic framework. Experimental validation is carried out using two sets of linear and nonlinear variable-speed simulation signals, one publicly available Ottawa bearing dataset, and one set of real onboard measurements. The results demonstrate that RE-ASDLA effectively suppresses background noise and reconstructs variable-speed fault features with high accuracy, even under low-SNR conditions. It maintains robust performance across different speed profiles and significantly enhances the adaptability of dictionary learning for variable-speed scenarios. Compared with fast path optimization and time-reassigned synchrosqueezing transform methods, RE-ASDLA shows superior performance in terms of reconstruction accuracy, time-frequency concentration, and fault feature representation.

Abstract:Aiming at the problem that the LIO-SAM algorithm in the SLAM algorithm of LiDAR lacks sufficient localization accuracy in complex environments, this paper proposes an improvement strategy focusing on two key aspects, feature point extraction and back-end point cloud matching. In the back-end matching, in view of its large inter-frame error fluctuation and poor robustness, this paper innovatively proposes an adaptive downsampling method based on pre-matching. The method effectively improves the initial matching accuracy through pre-matching operation, and dynamically adjusts the voxel filtering resolution based on the local density of the point cloud. Thus, the computational efficiency is significantly improved while ensuring the matching accuracy. In the front-end feature extraction, a two-stage feature filtering mechanism combining Early Cutoff and multi-scale voxel spatial covariance analysis is proposed to address the problems of redundant curvature computation, large sorting overhead, and low feature extraction rate of the near point cloud in LIO-SAM. The mechanism mainly focuses on the near point cloud: the redundant points are quickly eliminated by the local geometric change threshold, after which the covariance feature analysis is performed in the multiscale voxel grid. From this, the representative feature points with a balanced spatial distribution and a stable geometric structure are screened out, and the far point cloud is extracted by the original algorithm. Comparison experiments are carried out on the public dataset KITTI by selecting the stable sequence 07. Results show that, while the optimized algorithm provides only slight improvements in X- and Y-axis accuracy, it reduces the average absolute error of Z-axis by 26.44%, the RMSE by 24.43%, and the standard deviation by 30.24%. Furthermore, the algorithm has been deployed on a real-vehicle platform, where its robustness and engineering applicability have been verified.

Abstract:Although nondestructive testing technology is widely used in industrial field, it still faces great challenges in meeting the requirements of high-precision and high-resolution detection of weld defects. In particular, plate welds in critical sectors such as nuclear power and shipbuilding are often characterized by compact structural dimensions and complex spatial distributions, frequently located at corners, intersections of stiffeners, or other regions that are difficult to access. These features impose strict requirements on the size and placement of detection probes, which can typically only be deployed in close proximity to the weld. Existing ultrasonic inspection techniques generally rely on contact-based coupling or bulky transducers, making it difficult to simultaneously ensure excitation efficiency and defect detection sensitivity in confined spaces. Furthermore, these approaches often suffer from a low signal-to-noise ratio when detecting small-scale defects. To address these limitations, this study proposes a novel dual-arc V-shaped point-focusing electromagnetic acoustic transducer. The design employs a dual-arc meander-line coil combined with a symmetric V-shaped configuration, which not only confines the excited ultrasonic energy within a specific region but also significantly enhances the beam directivity and focusing performance. A finite element model was established to systematically investigate the influence of V-angle, central angle, and near-field characteristics on the transducer′s focusing capability and defect echo amplitude. The arc curvature of the dual-arc coil and the V-angle were optimized to improve the response to surface and near-surface micro-defects. Then, an experimental platform was built to detect the cracks and porosity defects in the weld. The research shows that the traditional transducer has limited detection effect on micro-cracks and hole-like defects, while the proposed new transducer can not only obtain clear reflected signals, but also achieve a minimum signal-to-noise ratio of 23.41 dB when detecting cracks with a size of 10 mm×0.5 mm×0.2 mm and blowholes with a diameter of 1 mm. This significantly improves the detection sensitivity and provides a high-sensitivity and non-contact detection scheme for weld defects detection of complex structures.

Abstract:To address the issues of limited datasets, tiny defect scales, and complex backgrounds in defect detection for CPCB in industrial settings, a CPCB defect detection method based on PXD-YOLO11s is proposed. Firstly, a high-quality dedicated dataset for CPCB defects is constructed, with a systematically designed acquisition scheme covering 14 types of typical defects such as deep scratches, sand holes, and open circuits. High-resolution defect images are captured using industrial cameras and professional optical equipment, and image preprocessing and data augmentation strategies are employed to enhance sample diversity and generalization. Secondly, improvements are made based on the YOLO11s network architecture: A parallel feature extraction module (ParNet) is introduced, which captures multi-scale feature information through a multi-branch convolutional structure while optimizing the convolution configuration to improve feature extraction efficiency; A dedicated small target detection layer (XsHead) is added to strengthen the recognition capability for tiny defects; the Soft-NMS mechanism is integrated in the post-processing stage to handle overlapping prediction boxes through confidence decay instead of direct suppression, effectively enhancing the detection performance for densely arranged defects; Finally, the loss function is adjusted by adopting DIoU loss to replace the traditional CIoU loss, enabling the model to focus more on the center distance and aspect ratio between prediction boxes and ground truth boxes, thereby improving target localization accuracy in complex backgrounds. Experimental results show that on the self-built CPCB dataset, the improved PXD-YOLO11s model achieves 5.2% and 8.5% increases in mAP50 and mAP(50~95) compared with the original YOLO11s, respectively. In addition, this method outperforms typical algorithms such as Faster R-CNN and YOLOv5s on the public PKU-Market-PCB dataset, demonstrating its excellent generalization ability and effective feature extraction performance. The proposed method significantly improves the CPCB defect detection accuracy and provides an efficient and reliable solution for intelligent industrial defect detection.

Abstract:To accurately detect foreign objects on coal mine belt conveyors under complex working conditions, a coal mine conveyor belt foreign object segmentation model based on improved DeepLabv3+ was constructed. Aiming at the difficulties in foreign object detection caused by interference factors such as high dust, uneven illumination, and mechanical vibration in coal mines, as well as practical requirements including the coexistence of multi-scale foreign objects and limited computing power of edge devices, the following improvements were made: The MobileNetv3 lightweight backbone network was introduced, and depthwise separable convolution was used to compress the computational load to 1/9 of that of traditional convolution. Meanwhile, the SE attention module was embedded to enhance high-frequency features such as edges and textures of foreign objects and suppress low-frequency channels corresponding to dust noise. The DASPP module was adopted to replace the traditional ASPP, and cross-layer feature dense interaction was realized by concatenating atrous convolution layers with different dilation rates, thereby improving the detection for multi-scale foreign objects. The ECANet channel attention mechanism was integrated, which enhanced feature expression ability through dimension-reduction-free global pooling and dynamic 1D convolution, further optimizing the distribution of feature weights. Experimental results show that the improved model achieved an average mean intersection over union of 87.1% and an F1-score of 86.7% on the CUMT-BelT dataset, with only 9.8 M parameters, 5.1 GFLOPs, and an inference speed of 38.6 fps. Compared with the original DeepLabv3+ model, the improved model increased accuracy by 4.6% and reduced computational load by 63.1%. In comparison with mainstream models such as PSPNet and U-Net, the improved model exhibits superior performance in key indicators including the missed detection rate of small-scale foreign objects, noise robustness, and adaptability to edge devices. This model provides a new approach to solving the problem of easy confusion between foreign objects and background features under complex working conditions. By simultaneously achieving high segmentation accuracy and low computational complexity, the proposed method supports real-time monitoring and contributes to advancing automation and intelligent development in the coal industry.

Abstract:Forward looking sonar plays an important role in underwater long-distance target detection and tracking. However, forward-looking sonar image sequences suffer from low frame rates and unclear target features, which can lead to target loss. In addition, effective tracking requires compensating for sonar carrier rotation and handling target occlusion to prevent trajectory discontinuities. To solve the above problems, the characteristics of sonar image sequences and target features were combined to improve existing tracking algorithms in this paper. To mitigate target loss in sonar tracking of ByteTrack algorithm, combined with the features of forward-looking sonar images, an improved correlation method was proposed. In the first correlation, a similarity measurement method based on Kalman filter was proposed by combining motion features and target appearance features, which improved the accuracy of tracking. To address rapid apparent target motion caused by sonar carrier rotation, rotation compensation was added to the ByteTrack algorithm using sonar attitude data to improve the accuracy of matching; Finally, the superiority of the improved association method and the combination of target appearance features were demonstrated through comparative experiments using similarity measurement algorithms. Compared with mainstream target tracking algorithms such as DeepSort, TransTrack, and ByteTrack, the improved model achieved a tracking accuracy of 76.8% and a tracking recall rate of 80.6%. Compared with the original ByteTrack algorithm, the improved ByteTrack has improved tracking accuracy by 9.4%, recall by 10.8%, and ID switching frequency by 46%. The fusion experimental results of detection and tracking show that the improved target detection tracking fusion algorithm has lower miss rate, lower false alarm rate, lower identity switching times, making it well-suited for underwater target detection and tracking with forward-looking sonar.

Abstract:Existing multimodal 3D object detection methods often rely on multi-scale spatial feature stacking, which leads to frequency information entanglement and limits detection accuracy. To address this issue, this paper proposes a Transformer-based method for feature frequency decoupling and fusion. Firstly, the input images are processed using discrete wavelet transform for multi-frequency decomposition. Separate high-frequency and low-frequency feature pyramids are constructed to capture detailed local textures and global structural semantics, respectively. Then, an asymmetric frequency update encoder is designed, where high-frequency features are treated as the primary components and updated through adaptive dynamic window encoding to enhance edge and texture representation. Meanwhile, sparse deformable attention is introduced to replace standard attention mechanisms for efficient low-frequency feature updating, enabling coordinated encoding across different frequency bands. A high-frequency guided voxel fusion module is further proposed, where multi-scale high-frequency features are projected into 3D voxel space via frustum-based mapping. Combined with an adaptive radius sampling strategy, this module effectively supplements the local structure of sparse point clouds and extracts critical voxel-level features. Finally, voxel and image features are unified in the bird′s-eye view space. A region-shift Transformer module is introduced to enhance cross-modal feature fusion using attention mechanisms. The proposed method is evaluated on the nuScenes, KITTI, and Waymo datasets. The method achieves 73.2% mAP and 74.3% NDS on the nuScenes test set, demonstrating strong performance in detecting small and distant objects. Moreover, real-vehicle experiments indicate that the method maintains high detection accuracy in complex and dynamic environments.

Abstract:Accurate three-dimensional shape measurement of vascular intervention surgical catheters is the key to improving surgical quality. The existing two-dimensional imaging, electromagnetic, and other methods have problems such as large measurement errors in the three-dimensional shape of catheters, which seriously restrict the safety and efficiency of surgery. This article proposes a shape reconstruction method for interventional catheters that integrates biplane perspective images and fiber Bragg grating sensing to meet the precise navigation requirements of vascular interventional surgical catheters, achieving accurate reconstruction of the three-dimensional shape of interventional surgical catheters. Firstly, based on the theory of fiber optic grating sensing, the curvature and directional angle of the grating nodes are solved. Meanwhile, to reduce the error accumulation caused by fiber optic torsion, the polar geometric stereo vision matching method is used to fuse the biplane perspective image to inverse solve the information of the conduit nodes, and the geometric parameters of the conduit curvature and directional angle reconstructed by the fiber optic grating are corrected. Finally, by combining the Frenet framework for iterative coordinate calculation, the corrected node geometric parameters are mapped to three-dimensional space to achieve accurate reconstruction of the three-dimensional shape of the interventional catheter. To evaluate the effectiveness and feasibility of the proposed fiber optic sensing interventional catheter shape reconstruction method by fusing biplane perspective images, an experimental test of catheter shape reconstruction is conducted. The results show that the proposed method can reduce the maximum reconstruction error of interventional catheter shape from 2.52 mm to 1.46 mm, which is about 42% lower than the traditional fiber optic grating shape reconstruction method. This indicates that the proposed method has broad application prospects in the precise navigation of vascular interventional surgical robots and the precise measurement of flexible robots.