Model Identification Research Articles

Reinforcement Learning for Efficient Identification of Soliton System Parameters Across Expansive Domains

Abstract Optical solitons in mode-locked fiber lasers and optical communication links have a wide range of applications. A crucial step in studying the transmission modes of optical solitons is to investigate the relationship between equation parameters and soliton evolution using deep learning techniques. However, current identification models have a limited parameter domain search range and are greatly influenced by initialization, often leading to divergence towards incorrect parameter values. This research harnesses reinforcement learning to revamp the iterative process of the parameter identification model. Through the development of a two-stage optimization strategy, the model is capable of conducting an accurate parameter search across arbitrary domains. The investigation encompasses a series of experiments on various standard and higher-order equations, illustrating that the innovative model overcomes the impact of initialization on the parameter search, and the identified parameters are guided towards their correct values. The enhanced model markedly improves experimental efficiency and holds significant promise for advancing the research of soliton propagation dynamics and addressing intricate scenarios.

A novel friction coefficient model parameters identification method for needle-tissue insertion procedure

Abstract In endoscopic liver vascular insertion surgeries, during the process of angiographic operation, the success of vascular staining depends on precise needle insertion control which heavily relies on experienced surgeons. Endoscopic vascular insertion surgical navigation system shows the potential to improve position precision, however it relies on needle-tissue interaction model and parameter identification to provide essential information for improving needle insertion accuracy, in which the friction coefficient is important parameters but difficult to determine. In this paper, a novel needle-tissue friction coefficient identification method was proposed with unknown tissue Young's modulus under endoscopic liver surgery scenarios. A modified friction coefficient model was proposed including the adhesion and elastic friction component to describe needle-tissue dynamic interaction process which can predict the friction coefficient more precisely. The proposed parameter estimation method based on the modified friction model can simultaneously estimate friction coefficient and Young's modulus. The proposed method was demonstrated by the friction coefficient measurement experiment. The results showed that the friction coefficient model prediction results agreed well with expected value. The proposed method can be applied to provide essential tissue-needle interactive information to improve needle insertion precision in the endoscopic liver vascular insertion surgery scenarios.

AIpollen: An Analytic Website for Pollen Identification Through Convolutional Neural Networks

With the rapid development of artificial intelligence, deep learning has been widely applied to complex tasks such as computer vision and natural language processing, demonstrating its outstanding performance. This study aims to exploit the high precision and efficiency of deep learning to develop a system for the identification of pollen. To this end, we constructed a dataset across 36 distinct genera. In terms of model selection, we employed a pre-trained ResNet34 network and fine-tuned its architecture to suit our specific task. For the optimization algorithm, we opted for the Adam optimizer and utilized the cross-entropy loss function. Additionally, we implemented ELU activation function, data augmentation, learning rate decay, and early stopping strategies to enhance the training efficiency and generalization capability of the model. After training for 203 epochs, our model achieved an accuracy of 97.01% on the test set and 99.89% on the training set. Further evaluation metrics, such as an F1 score of 95.9%, indicate that the model exhibits good balance and robustness across all categories. To facilitate the use of the model, we develop a user-friendly web interface. Users can upload images of pollen grains through the URL link provided in this article) and immediately receive predicted results of their genus names. Altogether, this study has successfully trained and validated a high-precision pollen grain identification model, providing a powerful tool for the identification of pollen.

YOLOX-S-TKECB: A Holstein Cow Identification Detection Algorithm

Accurate identification of individual cow identity is a prerequisite for the construction of digital farms and serves as the basis for optimized feeding, disease prevention and control, breed improvement, and product quality traceability. Currently, cow identification faces challenges such as poor recognition accuracy, large data volumes, weak model generalization ability, and low recognition speed. Therefore, this paper proposes a cow identification method based on YOLOX-S-TKECB. (1) Based on the characteristics of Holstein cows and their breeding practices, we constructed a real-time acquisition and preprocessing platform for two-dimensional Holstein cow images and built a cow identification model based on YOLOX-S-TKECB. (2) Transfer learning was introduced to improve the convergence speed and generalization ability of the cow identification model. (3) The CBAM attention mechanism module was added to enhance the model’s ability to extract features from cow torso patterns. (4) The alignment between the apriori frame and the target size was improved by optimizing the clustering algorithm and the multi-scale feature fusion method, thereby enhancing the performance of object detection at different scales. The experimental results demonstrate that, compared to the traditional YOLOX-S model, the improved model exhibits a 15.31% increase in mean average precision (mAP) and a 32-frame boost in frames per second (FPS). This validates the feasibility and effectiveness of the proposed YOLOX-S-TKECB-based cow identification algorithm, providing valuable technical support for the application of dairy cow identification in farms.

Identification Model of Traditional Village Cultural Landscape Elements and Its Application from the Perspective of Living Heritage—A Case Study of Chentian Village in Wuhan

At present, there are limitations in the cognition of the elements of traditional village cultural landscapes, less excavation of the elements of production-type cultural landscapes, little attention to the layering of cultural landscapes in different periods and the correlation between the elements of different types of cultural landscapes, and insufficient research on the living continuity of traditional village cultural landscapes. In the context of rural cultural revitalization, it is necessary to explore strategies for the conservation and development of traditional village cultural landscapes from the perspective of living heritage. By interpreting the core connotation of the concepts of living heritage and cultural landscape, a model of traditional village cultural landscape elements and a framework for the identification of living elements are constructed. On the above basis, Chentian Village in Wuhan is taken as an example to identify its living cultural landscape elements, and to analyze the evolution of its cultural landscape elements from the dimension of “time-function”. In Chentian Village, a total of seventeen cultural landscapes were identified. Of these, ten were cultural landscapes with functional continuity and living heritage characteristics; three were cultural landscapes whose existing functions have changed significantly from their initial functions; and four were cultural landscapes that were inactive and have lost their original functions. Finally, this study proposes corresponding protection and development strategies for living heritage and non-living heritage of cultural landscapes.

Construction of a Joint Newmark–Runout Model for Seismic Landslide Risk Identification: A Case Study in the Eastern Tibetan Plateau

The key to seismic landslide risk identification resides in the accurate evaluation of seismic landslide hazards. The traditional evaluation models for seismic landslide hazard seldom consider the landslide dynamic runout process, leading to an underestimation of seismic landslide hazard. Therefore, a joint Newmark–Runout model based on landslide dynamic runout is proposed. According to the evaluation results of static seismic landslide hazard, the landslide source points can be extracted, and the landslide dynamic runout process is simulated to obtain the dynamic seismic landslide hazard. Finally, the static and dynamic seismic landslide hazards are fused to obtain an optimized seismic landslide hazard. In September 2022, a strong Ms6.8 earthquake occurred in the eastern Tibetan Plateau, triggering thousands of landslides. Taking the 2022 Luding earthquake-induced landslide as a sample, the function relationship between seismic slope displacement and landslide occurrence probability is statistically modeled, which partly improves the traditional Newmark model. The optimized seismic landslide hazard evaluation of the Luding earthquake area is conducted, and then, the seismic landslide risk identification is completed by taking roads and buildings as hazard-affected bodies. The results show that the length of the roads facing very high and high seismic landslide risks are 3.36 km and 15.66 km, respectively, and the buildings on the Moxi platform near the epicenter are less vulnerable to seismic landslides. The research findings can furnish critical scientific and technological support for swift earthquake relief operations.

Addressing catastrophic forgetting in payload parameter identification using incremental ensemble learning

Collaborative robots (cobots) are increasingly integrated into Industry 4.0 dynamic manufacturing environments that require frequent system reconfiguration due to changes in cobot paths and payloads. This necessitates fast methods for identifying payload inertial parameters to compensate the cobot controller and ensure precise and safe operation. Our prior work used Incremental Ensemble Model (IEM) to identify payload parameters, eliminating the need for an excitation path and thus removing the separate identification step. However, this approach suffers from catastrophic forgetting. This paper introduces a novel incremental ensemble learning method that addresses the problem of catastrophic forgetting by adding a new weak learner to the ensemble model for each new training bag. Moreover, it proposes a new classification model that assists the ensemble model in identifying which weak learner provides the most accurate estimation for new input data. The proposed method incrementally updates the identification model while the cobot navigates any task path, maintaining accuracy on old weak learner even after updating with new data. Validation performed on the Franka Emika cobot showcases the model’s superior accuracy and adaptability, effectively eliminating the problem of catastrophic forgetting.

Identification of the Continuous-Time Hammerstein Models with Sparse Measurement Data Using Improved Marine Predators Algorithm

Identification of the Continuous-Time Hammerstein Models with Sparse Measurement Data Using Improved Marine Predators Algorithm

Dynamic model identification of the medieval bell tower of Casertavacchia (Italy)

The structural safeguard of tangible cultural heritage, monuments in particular, needs a multi-disciplinary approach involving historical–critical analysis, survey, in-situ and laboratory testing to provide a comprehensive knowledge on the object under investigation and build a structural model representative of the current state of the monument. In this paper, this approach is applied to the case study of the bell tower of San Michele Arcangelo cathedral in Casertavecchia (Italy). The historical–critical analysis was fundamental to understand the evolution of the tower over the centuries. Material tests were complemented by ambient vibration testing (AVT) which allowed for a complete characterisation of frequencies and mode shapes for the bell towers. A further model identification procedure was carried out to exploit the amount of information gathered within the documental and experimental activities and set up a reasonably accurate numerical model. The activities described in this paper highlight (a) the importance of correctly representing all substructures present in the monument, possibly hidden and built in different historical periods; (b) the possibility offered by model identification procedures to separate identifiable from less significant parameters, as well as selecting the best model in terms of representation complexity; (c) the effect of soil-structure and structure-structure interaction on the modal behaviour of the monument. In particular, the identification analyses without considering the inner concrete frame are consistently characterised by higher values of ωf and ωMAC compared to the analogous analyses in which all substructures were included. The best fit with the dynamic data is obtained by introducing deformable structure-soil connection while the effect of lateral constraints is hardly captured by the utilised setup.

Optimal experimental design for identification of hydrodynamic loading models

Abstract Hydrodynamic model-scale experiments are an intrinsic part of the design of marine structures, as they enable validating and calibrating the involved hydrodynamic numerical models. Such seakeeping experiments are generally conducted using a simple spring-based mooring system, with fixed properties throughout the tests. In the context of cyber-physical testing, the marine structure is kept in position in the laboratory by a virtual mooring system, the properties of which can be adjusted on the fly. This paper provides an algorithmic approach for maximizing the information gain from the test by optimizing the mooring stifness and damping parameters that define the experiment. The method is verified with synthetic data generated for the INO Windmoor 12MW floater, and is shown to be robust to noise and unmodeled effects.

An Analysis and Reliability-based Optimization Design Method of Trajectory Accuracy for Industrial Robots Considering Parametric Uncertainties

To address the challenges of poor trajectory accuracy in industrial robots, which has emerged as a technological bottleneck hindering further robots’ applications in high-precision manufacturing industries, this paper proposes a method for the analysis and reliability-based optimization design for industrial robots’ trajectory accuracy considering parametric uncertainties. Firstly, the dynamic equation of an articulated industrial robot with six degrees of freedom is derived, incorporating the Stribeck joint friction model, followed by the uncertain parameter identification of this dynamic model. Subsequently, an uncertainty simulation system for the robot is established based on the constructed dynamic model and the sensitivity of system uncertain parameters to the robot trajectory accuracy is analyzed, where 10 key parameters are obtained among 54 uncertain parameters. Finally, a reliability-based multi-objective optimization design methodology is proposed synthesizing the robot trajectory accuracy, manufacturing cost, and quality loss, to achieve tolerance design of the robot's parameters, and enables minimizing costs and quality losses while ensuring the robot's trajectory accuracy reliability. The performance and practicality of the proposed method were validated using a six-degree-of-freedom rotary joint serial industrial robot as an example.

Identification of heart rate dynamics during treadmill and cycle ergometer exercise: the role of model zeros and dead time

Background The response of heart rate to changes in exercise intensity is comprised of several dynamic modes with differing magnitudes and temporal characteristics. Investigations of empirical identification of dynamic models of heart rate showed that second-order models gave substantially and significantly better model fidelity compared to the first order case. In the present work, we aimed to reanalyse data from previous studies to more closely consider the effect of including a zero and a pure delay in the model. Methods This is a retrospective analysis of 22 treadmill (TM) and 54 cycle ergometer (CE) data sets from a total of 38 healthy participants. A linear, time-invariant plant model structure with up to two poles, a zero and a dead time is considered. Empirical estimation of the free parameters was performed using least-squares optimisation. The primary outcome measure is model fit, which is a normalised root-mean-square model error. Results A model comprising parallel connection of two first-order transfer functions, one with a dead time and one without, was found to give the highest fit (56.7 % for TM, 54.3 % for CE), whereby the non-delayed component appeared to merely capture initial transients in the data and the part with dead time likely represented the true dynamic response of heart rate to the excitation. In comparison, a simple first-order model without dead time gave substantially lower fit than the parallel model (50.2 % for TM, 47.9 % for CE). Conclusions This preliminary analysis points to a linear first-order system with dead time as being an appropriate model for heart rate response to exercise using treadmill and cycle ergometer modalities. In order to avoid biased estimates, it is vitally important that, prior to parameter estimation and validation, careful attention is paid to data preprocessing in order to eliminate transients and trends.

A fast deployable model for crack identification with laser thermography testing

This paper presents a novel feature extraction method, enabling efficient training and deployment of neural networks for rapid identification of crack defects in Laser Array Spot Thermography (LAST). We trained the crack defect identification model based on pixel-level features, encoding each pixel as a feature vector using Frangi filter, and classifying them using a neural network. Experimental results demonstrate that Frangi features are an effective method for distinguishing cracks, speckles, and background noise interference in the experiment. Furthermore, the model only requires a small region of interest (ROI) as training samples to achieve effective training and efficient crack identification under the same detection conditions, allowing for rapid deployment in practical inspections.

A multivariable model of clinical features for distinguishing sporotrichosis and Mycobacterium marinum cutaneous infection

To investigate differences in the clinical features of patients with sporotrichosis and cutaneous Mycobacterium marinum infection, and develop a prediction model for the initial identification. A total of 121 patients were selected based on the results from etiological culture. Logistic regression analyses were employed to identify the most valuable variables. In the multivariate logistic regression analysis, the following were finally considered independent predictors (OR = 3.650∼14.024, P<0.05) of cutaneous M. marinum infection: age<60 years (P = 0.031); female (P = 0.001); only the extremities involved (P = 0.031); clear triggers (P = 0.022); and presence of swelling (P = 0.024). A nomogram was developed based on these factors, and the area under the curve was 0.860. The model presented is based on clinical features and can help identify patients with sporotrichosis and cutaneous M. marinum infection. This may help increase the accuracy of doctors’ initial examination and diagnostic outcomes.

Formation of Civil Identity in the Educational Policy System in the Context of Transformation of the Political Space

The relevance lies in the institutionalization of education in the new subjects of the Russian Federation (DPR, LPR, Zaporizhia, Kherson regions, the Republic of Crimea, the federal city of Sevastopol), as well as the processes of re-socialization, adaptation to the socio-cultural, political, value-ideological realities of Russian society, which affects the construction of the civic identity of school youth. Understanding the processes of formation of civic identity in the educational policy system in the context of the transformation of the political space allows identifying the mechanisms of actualization and transformation of identity. Тhe experience of the post-Soviet states, as well as the processes taking place in Russia today, enables making interesting theoretical generalizations and practical recommendations.Aim and Tasks. Identification and systematization of practices, mechanisms and technologies for the formation of civil identity of schoolchildren in new subjects in the educational policy system of the Russian Federation, determination of existing problems and barriers, identification of models of civil identity formation in the context of comparing the experience of Russia and the EAEU countries.Methods. The study uses a focus group interview with senior and middle school students living in the new territories of the Russian Federation, an expert interview with representatives of institutional structures involved in the processes of political socialization and implementation of identity policy, as well as the analysis of secondary data reflecting the processes of formation of civil identity in the EAEU countries.Results. As a result, the key practices of forming the civil identity of schoolchildren in the new subjects of the Russian Federation are revealed and systematized, barriers and problems of implementing educational policy in the system of forming civil identity are identified. The authors propose recommendations for improving the model of formation of the identity of schoolchildren in the new subjects of the Russian Federation. The article shows the difference between the model of formation of Russian civil identity in the new regions and the models of formation of civil identity used in the EAEU countries.Conclusions. The study revealed the existing contradictions that arise in the process of adaptation to a new educational and value-ideological model, manifested in the assimilation of new social norms and educational standards; the difference in the level of political education and the degree of involvement in civic activity of young people was noted. It is necessary to search for effective practices and technologies that contribute to building a consistent model of civic identity formation in the basic education system of new subjects of the Russian Federation. The EAEU member states can use the practical experience of civil identity formation in the education system of new subjects of the Russian Federation to improve their own models of civil identity formation.

Nonlinear second order plus time delay model identification and nonlinear PID controller tuning based on extended linearization method

PID control systems based on the first order plus time delay model (FOPTD), which approximate the full system dynamics, are well-accepted for a wide range of linear processes. While such controllers can be applied to overdamped nonlinear processes, they often experience excessive overshoots and oscillations for general nonlinear processes. To overcome this limitation, we propose a novel method to design a nonlinear PID controller based on the second order plus time delay (SOPTD) model. The system nonlinearity requires parameter adjustments of the linearized model across operational ranges. Hence, in this work, it is handled by the extended linearization method (ELM), ensuring local stability under the assumptions of slow and small changes in operating points. Importantly, the model achieves global input-to-output stability even without the above constraints, provided there are no structural and parametric errors. The resulting nonlinear SOPTD model can describe changes in process gain and two time constants as the operation point varies. We demonstrate the applicability of our approach with a polymerization reactor simulation and liquid-level control experiments.

Dynamic modeling of post-combustion carbon capture process based on multi-gate mixture-of-experts incorporating dual-stage attention-based encoder-decoder network

Solvent-based post-combustion carbon capture (PCC) technology is a promising, near-term solution for decarbonizing power generation and industrial facilities. Model-based process simulation is crucial for the optimal design and operation of the PCC process. Recently, data-driven models have gained attention due to their adaptability, efficient computation and high accuracy. However, the nonlinearity, strong couplings and multi-time scale features of the PCC process pose significant challenges for model identification. To this end, this paper proposes a multi-gate mixture-of-experts incorporating dual-stage attention-based encoder-decoder (MMoE-DAED) network for dynamic modeling of the PCC process under wide operating conditions. An encoder-decoder composed of long short-term memory (LSTM) network is employed to extract features from the time-dependent input data and learn the complex dynamic interactions caused by the inertial and delay properties of the process. Dual-stage attention mechanism is incorporated into the encoder and decoder respectively to select the most relevant input features and their correlations within the time series data. To enhance multi-output prediction accuracy, multi-gate mixture-of-experts (MMoE) framework that considers correlations of multitask learning is implemented. Simulation results using operating data from a PCC experimental setup indicate that the proposed modeling approach accurately predicts the steady-state values and dynamic trends of the CO2 capture rate and stripper bottom temperature over a wide operating range. The RMSE, MAPE and R2 indices for the CO2 capture rate are 2.1592, 0.0295, 0.9641, respectively, and for the stripper bottom temperature are 0.1491, 0.0003, 0.9833, respectively. Validations on a PCC simulator further verify the accuracy and efficiency of the MMoE-DAED model, which enables an 80.87% reduction in computation time compared to the simulator. This paper points to a new direction for the data-driven dynamic modeling of complex energy conversion processes.

Identification of the Biomarkers for Chronic Gastritis with TCM Damp Phlegm Pattern by Using Tongue Coating Metabolomics.

This study aimed to establish a model for identifying chronic gastritis with the traditional Chinese medicine damp phlegm pattern by examining metabolite changes in the tongue coating of patients. It also explored the role of metabolic pathways in the pathogenesis of this condition. This cross-sectional study involved 300 patients diagnosed with chronic gastritis. Of these, 200 patients exhibited the damp phlegm pattern, while 100 did not. Metabolomic methods employing GC-TOF-MS and UHPLC-QE-MS were utilized to identify various metabolites in the tongue coating of patients. An identification model for chronic gastritis with the damp phlegm pattern was created based on ROC curves derived from differential biomarkers. Additionally, 50 samples not included in model construction were collected for external validation. Comparison of the damp phlegm pattern group with the non-damp phlegm pattern group revealed a total of 116 differential metabolites. Among these, lipids and lipid-like compounds were most abundant, comprising 27 types, which included four lipid metabolites related to sphingomyelin metabolism. The ROC model, which included phenol, 2.6-diaminoheptanedioic acid, and N-hexadecanoyl pyrrolidine, demonstrated the highest accuracy, with accuracy, sensitivity, and specificity metrics of 94.0%, 91.0%, and 87.0%, respectively. Furthermore, external validation using tongue coating metabolites from 50 patients revealed accuracy, sensitivity, and specificity in the validation set of 93.9%, 90.6%, and 83.3%, respectively. Differential metabolites between patients with the damp phlegm pattern and those without are primarily lipids and lipid-like compounds. N-hexadecanoyl pyrrolidine, phenol, and 2.6-diaminoheptanedioic acid may serve as potential biomarkers for chronic gastritis characterized by the damp phlegm pattern.

A novel multi-information fusion CNN for defect detection in laser soldering of SAC305

Identification of laser soldering of lead-free solder Sn-3.0Ag-0.5Cu (SAC305) in electronic packaging was still an enormous challenge. It was difficult to detect defects in large-scale production. This work proposed an identification model based on multi-information fusion convolutional neural network (MIFCNN) for inspecting laser soldering process. In this method, the forty images in chronological order and the temperature data were combined as the input to be utilized in detecting defects. The results demonstrated that MIFCNN had best accuracy for three types of joints with accuracy of 98.28 % due to the combination of images and temperature information. The ICNN and TCNN had poor recognition accuracy for the warpage defect with 73.9 % and the poor wetting defect with 66.5 %, respectively. This was because the images and temperature information were the key to identifying the poor wetting defects and warpage defects, respectively. The poor wetting defect could be recognized by difference of contact angle, while the warpage defect could be significantly detected by maximum temperature. This work could help detecting defects of laser soldering in the actual production and widen the application of MIFCNN in the field of laser soldering.

Identification and validation of a prognostic model based on immune-related genes in ovarian carcinoma.

A novel valuable prognostic model has been developed on the basis of immune-related genes (IRGs), which could be used to estimate overall survival (OS) in ovarian cancer (OC) patients in The Cancer Genome Atlas (TCGA) dataset and the International Cancer Genome Consortium (ICGC) dataset. This prognostic model was engineered by employing LASSO regression in training cohort (TCGA dataset). The corresponding growth predictive values of this model for individualized survival was evaluated using survival analysis, receiver operating characteristic curve (ROC curve), and risk curve analysis. Combined with clinical characteristics, a model risk score nomogram for OS was well built. Thereafter, depended on the model risk score, patients were divided into high and low risk subgroups. The survival difference between these subgroups was measured using Kaplan-Meier survival method. In addition, correlations containing pathway enrichment, treatment, immune cell infiltration and the prognostic model were also analyzed. We established the ovarian cancer cell line W038 for this study and identified the performances of GBP1P1 knockdown on a series of activities including cellular proliferation, apoptosis, migration, and invasion of W038 cells in vitro. We constructed a 25-genes prognostic model (TNFAIP8L3, PI3, TMEM181, GBP1P1 (LOC400759), STX18, KIF26B, MRPS11, CACNA1C, PACSIN3, GMPR, MANF, PYGB, SNRPA1, ST7L, ZBP1, BMPR1B-DT, STAC2, LINC02585, LYPD6, NSG1, ACOT13, FAM120B, LEFTY1, SULT1A2, FZD3). The areas under the curves (AUC) of 1, 2 and 3 years were 0.806, 0.773 and 0.762, in the TCGA cohort, respectively. Besides, the effectiveness of the model was verified using ICGC testing data. Univariate and multivariate Cox regression analysis exposes the risk score as an independent prognosis predictor for OS both in the TCGA and ICGC cohort. In summary, we utilized comprehensive bioinformatics analysis to build an effective prognostic gene model for OC patients. These bioinformatic results suggested that GBP1P1 could act as a novel biomarker for OC. GBP1P1 knockdown substantially inhibited the proliferation, migration, and invasion of W038 cells in vitro, and increased the percentage of apoptotic W038 cells. The analyses of genetic status of patients with 25-genes model might improve the ability to predict the prognosis of patients with OC and help to select patients suit able to therapies. Immune-related gene GBP1P1 might serve as prognostic biomarker for OC.