搜索结果：Verified

In this paper, an improved state-space method (SSM) is proposed for gait feature extraction. By introducing zero-phase component analysis Whitening (ZCA Whitening) and an algorithm to search estimated echo as the preprocessing method, pedestrian echoes are divided into three groups according to the frequency probability density: torso, feet, and other segments. Two channels of echoes are selected as inputs to the SSM, which is employed to identify the corresponding micro-Doppler trajectory. On this basis, five gait features of torso amplitude, stride length, walking cycle, torso maximum speed, and feet maximum speed are extracted. Simulation based on the Boulic model, compared with the traditional SSM, demonstrated that there is no need to estimate the model order and that a more accurate torso micro-Doppler trajectory and effective micro-motion features of the feet can be obtained by the proposed method. Finally, 77 GHz FMCW radar was used to collect the echoes of four pedestrians. The classifier was designed based on a support vector machine (SVM), and the classification experiment verified the effectiveness of the extracted gait features.

Fatty acid degradation (FAD) plays a crucial role in maintaining cellular energy homeostasis, with its dysfunction serves as an important pathological basis for the progression of various diseases. However, the specific regulatory mechanisms of this process in osteoarthritis (OA) remain to be further elucidated. This study aims to identify potential FAD-associated biomarkers and to investigate the role and potential mechanisms of FAD in OA. OA-related datasets and FAD-associated genes were retrieved from publicly accessible databases. Multiple bioinformatics methods were employed to reveal the potential connections among the aforementioned genes. Screening FAD-associated differentially expressed genes highly correlated with OA (hub OA-FADEGs) using machine learning methods. Single-sample gene set enrichment analysis (ssGSEA) was employed to characterize immune cell infiltration in OA and to explore their correlations with FADEGs. Additionally, scatter plots were used to evaluate the diagnostic efficacy of hub OA-FADEGs. Finally, enrichment analysis of hub OA-FADEGs and their corresponding therapeutic drugs was performed using the Drug Signatures Database (DSigDB). Machine learning algorithms were applied to screen for hub OA-FADEGs, identifying APOD, COL1A1, SULF1, CHI3L1, PENK, and ADM as genes that are significantly upregulated or downregulated in OA samples. These results were subsequently verified by qRT-PCR. Furthermore, the aforementioned genes all exhibit strong diagnostic efficacy for OA. Ultimately, we identified 28 therapeutic drugs that may target hub OA-FADEGs using DSigDB. Based on comprehensive bioinformatics analysis, this study proposes that 6 key hub OA-FADEGs, including APOD, COL1A1, SULF1, CHI3L1, PENK, and ADM, could serve as potential diagnostic biomarkers for OA and highlights their regulatory roles in disease progression. These findings provide novel insights into the metabolic pathogenesis underlying OA.

Spinal fusion is widely performed to treat instability of various spinal pathologies. Currently, the clinical gold standard is still autografts, which unfortunately are limited by donor site morbidity, restricted supply, and inconsistent outcomes. To overcome these challenges, we developed an organic-inorganic nanohybrid (click-ON) cement based on poly(propylene fumarate) (PPF) polymers cross-linked through the strain-promoted azide-alkyne cycloaddition (SPAAC) bioorthogonal click chemistry. This catalyst-free system cures rapidly in situ without external energy or toxic initiators, enabling safe and practical surgical handling. The cement was further reinforced with osteogenic nanohydroxyapatite (nHA) to improve osteoinductivity and microspheres releasing bioactive recombinant human bone morphogenetic protein 2 (rhBMP-2) and recombinant human vascular endothelial growth factor (rhVEGF) to promote coupled osteogenesis and angiogenesis. In a sheep lumbar spinal fusion model, the click-ON cement was implanted with an injectable formulation in the interbody space and a moldable formulation in the posterolateral fusion site with clinically used autograft/rhBMP-2 as the positive control. Longitudinal CT imaging demonstrated fusion with bridging bone formation across both the interbody space and posterolateral region after 6 months. Histological analyses confirmed extensive new bone deposition, integration with host tissue, and vascular ingrowth within the cement, while immunohistochemical staining showed the colocalization of CD31 and alkaline phosphatase (ALP), indicating active angiogenesis and osteogenesis, respectively. The outcomes are comparable to the positive control, which are clinical gold standard bone grafts. Manual palpation further verified the mechanical stability of the fused segments in sheep with a bioactive click-ON cement. These results established click-chemistry-enabled PPF-based cement as a promising alternative to autografts, offering advantages in moldability, biological activity, and functional fusion outcomes.

Metamaterials possess high freedom on structural design, yet their ability to modulate electromagnetic waves is subject to intrinsic constraints that are independent of specific meta-atom geometries. The constraints are revealed by analyzing the statistical amplitudes and phases of transmission and reflection wave in some representative metamaterials. Based on scattering theory, a reconstructed and more general description of the electromagnetic modulation process in metamaterials is established. Two explicit and geometry-independent corollaries concerning the coupling between transmission and reflection waves are further obtained and verified. The results provide a new perspective on the fundamental modulation mechanism of metamaterials on electromagnetic waves.

This study analyzed factors associated with complete cytoreductive surgery and postoperative frailty in advanced ovarian cancer using key neutrophil extracellular trap (NETosis) markers-neutrophil elastase (NE), myeloperoxidase (MPO), and citrullinated histone H3 (Cit-H3). A risk prediction model was developed and validated. In this prospective cohort study, 189 advanced ovarian cancer patients (2020-2023) were classified into frail (n=41) and non-frail (n=148) groups based on postoperative status, and all patients were followed up for 2 years. Clinical data were collected, and risk factors for postoperative frailty in advanced ovarian cancer were identified using a machine learning method (LASSO - XGBoost). A nomogram-based prediction model was constructed. Internal validation and decision curve analysis confirmed favorable predictive efficacy and clinical net benefit of the model. Significant differences were found between groups in age, education, marital status, daily activity, nutrition score, State-Trait Anxiety Inventory (STAI), Pittsburgh Sleep Quality Index (PSQI), NE, MPO, and Cit-H3 (P<0. 05). Kaplan-Meier survival curves showed that postoperative frailty was associated with worse prognosis (P<0. 05). Eight common risk factors were identified through overlapping screening by two machine learning methods, LASSO regression and XGBoost. Multivariate Logistic regression confirmed age, STAI, MPO, NE, and Cit-H3 as independent risk factors (all OR>1, P<0. 05), while nutrition score was protective (OR<1, P<0. 05). The constructed nomogram model exhibited good discriminative ability (AUC = 0. 882) and calibration (C-index=0. 856, calibration slope=0. 92). The Hosmer-Lemeshow test indicated good model fit (P = 0. 893), and decision curve analysis demonstrated high net clinical benefit. Postoperative frailty in advanced ovarian cancer is associated with a multifactorial profile, primarily driven by age, nutritional status, STAI scores, MPO, NE, and Cit-H3 levels. The nomogram model was constructed based on these factors initially demonstrated favorable predictive efficacy, and it is expected to serve as an auxiliary tool for the early clinical identification of high-risk populations. However, this study is an exploratory single-center, small-sample research, and its conclusions still need to be further verified through external validation studies with multicenter and large-sample designs.

China faces a severe imbalance between the supply and demand of formal childcare services for infants and toddlers aged 0-3, with rural research and resource allocation falling far behind urban areas. This study aimed to identify key determinants of rural childcare demand in western China's less-developed regions and to screen the optimal predictive machine learning (ML) prediction model, based on Andersen's Behavioral Model. A cross-sectional survey with purposive and multi-stage sampling was conducted in southwestern Guizhou Province, collecting valid data from 1,116 rural families with infants and toddlers. Seven ML algorithms were applied to construct childcare demand prediction models with comprehensive model evaluation, SHAP analysis for feature identification, and threshold analysis for key factors performed subsequently. The random forest (RF) model was identified as the optimal model, demonstrating robust generalization and discriminative ability. SHAP analysis revealed childcare flexibility, overall childcare quality, early education, and teacher professionalism as the four core positive determinants of rural childcare demand. Threshold analysis further defined the optimal critical values of these key factors and verified their strong practical predictive performance, with childcare flexibility and overall quality showing the most prominent effects. This study confirms that machine learning can effectively identify the determinants of rural childcare demand. The four service-related factors are the most influential drivers. The findings provide empirical evidence for optimizing rural childcare services and formulating demand-oriented childcare policies to promote the high-quality and inclusive development of rural childcare systems in less-developed regions in China.

To evaluate the potential clinical effect of Yanghe decoction on diabetic foot. We comprehensively searched Web of Science, PubMed, The Cochrane Library, Embase, CNKI, Wanfang, and VIP databases from their inception to December 6, 2025, to identify randomized controlled trials (RCTs) investigating Yanghe decoction for DF. Meta-analysis was performed using RevMan 5.4 and Stata 15.0 software. A total of 15 studies with 1224 patients were included, 614 in the treatment group and 610 in the control group. The results suggest that the total effective rate of the treatment group may be higher than that of the control group (RR&#x2009;=&#x2009;1.23, 95% CI: 1.17 to 1.29, p < 0.00001). The wound healing time of the treatment group may be shorter than that of the control group (MD&#x2009;=&#x2009;-9.36, 95% CI: -14.21 to -4.50, p = 0.0002), and the treatment group may have potential advantages over the control group in the improvement of wound area after treatment (SMD&#x2009;=&#x2009;-2.57, 95%CI: -3.94 to -1.20, p = 0.0002). There was no statistically significant difference in the ankle-brachial index (ABI) between the treatment group and the control group (MD&#x2009;=&#x2009;0.05, 95% CI: -0.09 to 0.18, p = 0.51). The funnel plot of the total effective rate between the treatment group and the control group suggested that there was a publication bias in the effective rate. Further analysis of the results by the nonparametric trim and filling method suggested that the results of the meta-analysis were relatively stable, and the possible publication bias did not substantially affect the results. Low certainty evidence suggests that Yanghe decoction may potentially improve the clinical symptoms of diabetic foot, may shorten the wound healing time, and may reduce the wound area. However, the independent clinical efficacy of Yanghe decoction for diabetic foot cannot be determined due to the high heterogeneity of cointerventions in the included studies. This study is further limited by the insufficient quantity and quality of the included studies, and the above tentative implications need to be verified by more large-sample and multicenter RCT studies that standardize interventions and control for cointervention variables.

The organizational growth and sustainability are closely related to how well the organizational system is functioning. In particular, the organizational innovation and competitiveness are related to the creativity and innovation through knowledge sharing among organizational members in a competitive situation. In relation to such background, this study focused on the learning organization system as a key variable that drives to knowledge sharing among organizational members. As a product of the knowledge economy era, the learning organization system plays a crucial role in an organization's development process. The establishment of a learning organization system helps create a shared learning platform, which reduces the cost of acquiring knowledge for employees, promotes joint learning among employees, and effectively facilitates the rapid dissemination of knowledge within the organization. This practice effectively stimulates employees' innovation ability and enhances their innovation performance, which, in turn, provides important support for the sustainable development of organizations. Therefore, this study argues that organizational learning systems are closely related to employees' innovation performance. In this context, this study investigates whether learning organizational systems can improve employees' innovation performance and the mediating role of employees' knowledge sharing behaviors in promoting innovation performance. Most previous studies focused only on the mediating or moderating roles of the model. This study expands the research field by examining the moderating role of organizational cultural identification. We also verified its moderating role. Overall, this study aims to determine improvements in the innovation performance of employees under a learning organization system. Based on this, suggestions are proposed to strengthen and optimize the learning organization system and provide a theoretical basis for related research. This study collected survey data from 300 Chinese SME employees for empirical analysis. The results show that the learning organization system has a positive effect on employee innovation performance. Knowledge sharing can also positively influence innovation performance. In addition, the interaction between organizational cultural identification and the learning organizational system improves employees' knowledge sharing. Overall, we contributed to the field of learning organization system research and fill the theoretical and practical gaps in existing related research.

Microelectrode arrays (MEAs) are essential tools for recording and stimulating electrogenic tissues, but their fabrication typically depends on complex, costly, and mask-based cleanroom processes. While inkjet-printed MEAs have increasingly been explored as low-cost alternatives, most demonstrations have focused on cardiac cell recordings, with only a limited number of studies showing neuronal recordings. Furthermore, no work to date has demonstrated neuronal interfacing, combining single-unit recording with electrical stimulation, using inkjet-printed MEAs. Here, we investigate whether inkjet-printed MEAs enable both extracellular single-unit neuronal recording and reliable electrical stimulation. We fabricated gold microelectrodes on flexible foils via maskless inkjet-printing, insulated them with printed SU-8 (an epoxy-based dielectric), and characterized their morphology using scanning electron microscopy, atomic force microscopy, and profilometry, and their electrochemical behavior using impedance spectroscopy and cyclic voltammetry. The printed gold formed a rough nanoparticle-based morphology, resulting in an increased effective electrochemical surface area. This morphology enabled low electrode impedances and high charge injection during voltage-controlled stimulation. We assessed functional performance in ex vivo retinal tissue. The inkjet-printed MEAs enabled reliable single-unit recordings with signal-to-noise ratios comparable to cleanroom-fabricated commercial devices and cell activation upon electrical stimulation with biphasic pulses. The electrodes were reusable and noncytotoxic, verified via a standard cell viability assay. These results establish the first inkjet-printed microelectrodes capable of neuronal interfacing, demonstrating that printed MEAs can match the functional performance of conventional microfabricated devices. This work positions inkjet-printing as a scalable, easily adaptable, low-cost manufacturing technique for flexible MEAs with rough gold electrodes suitable for neurotechnology applications.

[Purpose] This case series aimed to evaluate the effects of a standardized exercise dosage using a cycle ergometer on body weight and patient-reported outcome measures among overweight female patients with knee osteoarthritis. [Participants and Methods] We conducted a two-phase study. Exercise dosage was standardized using kilocalories per kilogram of body weight per week. Study 1 verified the feasibility and safety of progressive dosage in a single participant, while Study 2 used an ABCB-type single-case design involving three participants. The participants performed a cycle ergometer exercise with standard physical therapy. Body weight was the primary outcome, and the Japan Knee Osteoarthritis Measure was the secondary outcome. [Results] In Study 1, the participant safely achieved an energy expenditure of up to 10 kilocalories per kilogram of body weight per week; however, symptoms resembling knee buckling occurred at an expenditure of 11 kilocalories per kilogram of body weight per week. In Study 2, no adverse events were observed. One participant significantly reduced body weight during Phase B2, and all participants exhibited an improvement in Japan Knee Osteoarthritis Measure scores. [Conclusion] Standardized exercise dosage based on kilocalories per kilogram of body weight per week provides a safe and effective method for overweight female patients with knee osteoarthritis.

Membrane potential is an observable output, whereas the ion channel pathway governs the internal current partitioning that shapes neural firing modes. Here, we construct a memristive FitzHugh-Nagumo neural circuit in which a tunable diversion branch is connected to the canonical ion channel branch, enabling the regulation of neural firing through the controlled redistribution of channel current. Physical implementation of this strategy is verified by incorporating different electric elements in the sub-branch circuit: (i) a shunting capacitor for differential-type diversion and (ii) a shunting inductor in series with a protective resistor for integral-type diversion. For the two controlled circuits, physical equations and field energy functions are derived, and the corresponding theoretical models and corresponding energy functions are obtained and further checked by the Helmholtz theorem. Numerical analysis shows that capacitive shunting can trigger an abrupt collapse of the inductive energy level in L1 and thereby induce firing-mode transitions, whereas inductive shunting produces markedly weaker modulation over comparable parameter ranges. The model also exhibits noise-induced stochastic resonance, and an adaptive energy-guided regulation law controls the electrical activities effectively. The results suggest that control of firing patterns depends on the physical property of the shunting element and provide a physically interpretable strategy for ion-channel-level regulation in memristive neural circuits.

As the core power unit of complex electromechanical systems, accurate health assessment of diesel engines is essential for safe operation. The Interval Belief Rule Base (IBRB) method integrates observed data with expert knowledge to support system assessment. However, engine operating parameters change over time because of wear and aging. Additionally, traditional optimization methods struggle to balance global search speed with local convergence efficiency. To address these issues, this paper proposes an Interval Belief Rule Base method based on Hybrid Optimization and Adaptive Intervals (IBRB-HOAI). First, an adaptive reference interval is introduced by combining K-means clustering and quantile interval estimation, dynamically generated based on the actual operating state of the engine. The health assessment baseline is optimized. The applicability of the model is enhanced. Second, the global exploration ability of particle swarm optimization is combined with the local refinement ability of the projected covariance matrix adaptation evolution strategy. The model parameters are collaboratively optimized. Finally, experimental verification is conducted on a diesel engine dataset containing 2700 sample points. Compared with the traditional IBRB method, the proposed method achieves a significant reduction in MSE of 97.5%. It outperforms other machine learning methods. The effectiveness of the proposed method is verified.

Memristor-based neuromorphic computing offers a promising pathway for efficient in-memory processing. However, the scalability and reliability of such systems are severely compromised by parasitic resistances (including line and input resistances) in crossbar arrays, which cause significant IR-drop during vector-matrix multiplication (VMM). Existing research often suffers from high computational latency or relies on the precise extraction of parasitic parameters, which is impractical and computationally expensive for large-scale integration. To overcome these limitations, we propose a Parameter-Agnostic Adaptive Compensation (PAAC) method based on a distributed linear approximation model. By analyzing the circuit characteristics, we conquered the challenge of coupling between parasitic effects and output current, deriving a simplified linear relationship that requires no prior knowledge of specific resistance values. The PAAC method involves only a single-step pre-calibration experiment to determine a global compensation factor, achieving an ultra-low computational complexity during inference. We validated the method using a comprehensive two-stage strategy: board-level hardware experiments confirmed its feasibility by reducing current distortion from 71% to 2%, while extensive large-scale HSPICE simulations verified its scalability, restoring classification accuracy from 89% to 95%. This work provides a robust, low-overhead solution that eliminates the dependency on precise parameter modeling, facilitating the realization of large-scale, high-precision neuromorphic hardware.

This dataset compiles attacks on energy infrastructure in Nigeria between 2009 and 2025. It integrates geospatially attributed incident-level data with aggregated estimates of physical and economic losses. The dataset covers deliberate disruptions to oil and gas infrastructure and electricity transmission systems across Nigeria's six geopolitical zones. It consists of two complementary components. The first is a geo-referenced incident data (2011-2025) consisting of 161 recorded attacks, with information on date, location, actor type, targeted asset, and attack method. The second is an annual volume-value dataset (2009-2025) reporting crude oil losses and associated revenue estimates resulting from pipeline vandalism, crude theft, and spill events. Data were compiled from verified open sources, including the Nigeria Security Tracker (Council for Foreign Relations), domestic and international media reports, and industry and regulatory publications. These data were processed using reproducible filtering and classification scripts. Together, these datasets enable quantitative assessment of the physical scale and economic implications of crude oil theft in Nigeria. They also support comparative, temporal, and scenario-based analyses across research, policy, and public-interest applications. All data and supporting files are openly available in a Zenodo repository.

This study aimed to investigate the role of lysine acetyltransferase 6A (KAT6A) in lung cancer progression and its potential involvement in Nrf2-related oxidative stress regulation. KAT6A was overexpressed or silenced in A549 and H1299 lung cancer cells. KAT6A expression was verified by quantitative reverse transcription polymerase chain reaction and Western blot. Cell Counting Kit-8 and Transwell assays showed that KAT6A overexpression promoted cell proliferation and invasion, whereas KAT6A silencing suppressed cell proliferation. KAT6A overexpression decreased the expression of Keap1 protein and enhanced Nrf2 signaling activity. Oxidative stress evaluation using 2',7'-dichlorodihydrofluorescein diacetate staining, malondialdehyde (MDA) detection, and superoxide dismutase (SOD) activity assays indicated decreased reactive oxygen species levels, reduced MDA content, and elevated SOD activity. Co-immunoprecipitation confirmed the interaction between KAT6A and Nrf2, and dual-luciferase reporter assays showed enhanced Nrf2 transcriptional activity on the heme oxygenase-1 promoter. Silencing Nrf2 reversed the effects of KAT6A on proliferation. Immunohistochemistry of clinical lung adenocarcinoma samples showed&#xa0;that high KAT6A expression correlated with advanced tumor stage and shorter overall survival. These findings suggest that KAT6A regulates oxidative stress via the Keap1-Nrf2 pathway, thereby promoting malignant progression in lung adenocarcinoma, and may serve as a potential prognostic biomarker.

Odometry estimation, which calculates the trajectory of a moving object across timeframes, is a critical and time-consuming function in SLAM (Simultaneous Localization and Mapping) systems. Although LiDAR-based sensing is most popular for outdoor and long-range applications because of its ranging accuracy, the sparsity of laser point cloud poses a significant challenge to feature extraction and matching in odometry estimation. In this paper, we investigate odometry estimation from two aspects, i.e., algorithm optimization, and system design/implementation. In algorithm optimization, we present an image feature-assisted odometry estimation scheme that leverages the richness of image information captured by a companion camera to enhance the accuracy of laser point cloud matching. This also serves as a screening mechanism to reduce the matching size and lower the computing complexity for a higher estimation rate. In addition, various schemes, such as adaptive threshold in image feature point selection, principal component analysis (PCA)-based plane fitting for laser point interpolation, and Gauss-Newton optimization for calculating the transform matrix, are also employed to improve the accuracy of odometry estimation. The performance of improved odometry estimation is verified using an existing FLOAM (Fast Lidar Odometry and Mapping) framework. The KITTI dataset for autonomous vehicles with ground truth was used as the test bench. Simulation results indicate that the translation error and rotation error can be reduced by 16.6% and 1.3%, respectively. Computing complexity, measured as the software execution time, also reduced by 63%. In system implementation, a hardware/software (HW/SW) co-design strategy was adopted, where complexity profiling was first conducted to determine the task partitioning and time-consuming tasks are offloaded to a hardware accelerator. This facilitates real-time execution on a resource-constrained embedded platform consisting of a microprocessor module (Raspberry Pi) and an attached FPGA board (Pynq Z2). Efficient hardware designs for customized DSP functions (adaptive threshold and PCA) were developed in an FPGA capable of completing one data frame in 20ms. The final system implementation met the target throughput of 10 estimations per second, and can be scaled up further.

To develop a fully green and non-toxic wood adhesive with improved water resistance and bonding performance for soybean meal (Glycine max (L.) Merr.)-based adhesives, oxidized tannin (OTN) was obtained by the laccase treatment of waxberry tannin (TN), a natural polyphenolic polymer, and then blended with soybean meal (SM) to prepare an oxidized tannin-soybean meal adhesive (OTS). Laccase-mediated oxidation converted the tannin polymer into quinone-rich oxidized polymeric structures, which reacted with amino groups in soybean meal proteins through Michael addition and Schiff base reactions to form a covalently crosslinked polymeric network. Under the optimal conditions of a laccase dosage of 10%, an oxidation time of 6 h, an OTN:SM mass ratio of 0.5:1, and a hot-pressing temperature of 160 &#xb0;C, plywood bonded with OTS exhibited a wet shear strength of 0.85 MPa at 63 &#xb0;C, representing a 136% increase over that of the neat soybean meal adhesive, and showed slightly higher bonding performance than the commercial urea-formaldehyde (UF) resin under boiling-water conditions. Structural analyses (FT-IR and XPS) verified quinone formation and carbon-nitrogen single and double bonds. Thermal analyses (DSC and TGA) revealed improved curing reactivity and significantly enhanced thermal stability compared with the neat soybean meal adhesive.

A new molecular beam calibration facility, the MOlecular Beam for Instrument Utilization and Simulation (MOBIUS), has been developed to enable realistic, in situ testing and calibration of spaceborne neutral mass spectrometers. The facility is based on a differentially pumped beamline featuring a high-pressure, high-temperature nozzle source and modular diagnostic capabilities. MOBIUS is specifically engineered to reproduce the high-speed, directed neutral particle fluxes encountered by spacecraft in planetary exospheres and upper atmospheres. We demonstrate the system's performance by characterizing supersonic beams generated across multiple gases. When operating with seeded gas mixtures (hydrogen-argon), the facility reliably produces neutral beams in the 3-5&#xa0;km&#xa0;s-1 range, successfully achieving the high relative velocities required for instruments such as the Strofio mass spectrometer onboard BepiColombo. Furthermore, neutral flux is verified to be on the order of 1015&#xa0;cm-2&#xa0;s-1, and the resulting beam profiles are confirmed to be uniform and well-collimated. These results establish MOBIUS as a complete and necessary testbed for accurately characterizing the response of instruments under relevant in-orbit conditions.

This paper presents an adaptive noise reduction method based on Complementary Ensemble Empirical Mode Decomposition (CEEMD) to address the non-stationary characteristics and noise interference present in friction vibration signals from mechanical equipment. and friction testing machine simulation experiments. The performance of CEEMD and Ensemble Empirical Mode Decomposition (EEMD) was compared through MATLAB R2023b simulations and experiments conducted on a friction testing machine. CEEMD achieved a computational efficiency 85.6% higher than that of EEMD and effectively reduced mode aliasing. Among them, the adaptive correlation coefficient screening method performed well in signal reconstruction, and the high correlation (correlation coefficient > 0.8) between the denoised signal and the laboratory noise signal was verified using the multi-scale permutation entropy (MPE) theory, which is of great significance for early diagnosis of mechanical faults, prediction of equipment life and timely maintenance decisions.