Search

Accurately predicting methane adsorption capacity in coal is crucial for assessing coalbed methane resources and ensuring safe extraction. Conventional methane isotherm adsorption experiments often suffer from human error and experimental artifacts, leading to inaccurate and poorly reproducible outcomes. Furthermore, they are time-consuming to conduct, requiring specific and well calibrated experimental equipment. In this paper, a Random Forest (RF) algorithm is introduced to improve the accuracy and reliability of methane adsorption capacity prediction. Approximately 200 sets of experimental data, including parameters such as experimental temperature, equilibrium pressure, moisture, ash content, and volatile matter of coal samples, were collected and analyzed to establish a prediction model based on the RF algorithm. The robustness and reliability of the model were validated using K-Fold cross-validation and hyperparameter optimization. The results indicate that the Random Forest algorithm performs exceptionally well in predicting methane adsorption capacity, with optimal values for mean squared error (MSE) and the coefficient of determination (R2), demonstrating a high correlation between predicted and actual values. Machine learning algorithms are innovatively combined with traditional experimental methods in this study. By training the model using large data sets, issues of error and reproducibility in traditional experiments are addressed, improving experimental efficiency and providing a more reliable method for evaluating coalbed methane resources. To some extent, the method can replace traditional methane isotherm adsorption experiments in coal, improving prediction accuracy and efficiency and demonstrating promising prospects for wide application.

Polarization navigation is an emerging autonomous navigation technology suitable for unmanned underwater vehicles (UUVs). Nonetheless, light refraction poses challenges for underwater polarization navigation as it alters the direction and amplitude of the polarization electric vector (E-vector). In this article, we propose a novel underwater heading determination method with inertial navigation system/polarization sensor integration in consideration of light refraction. In view of light deflection and energy attenuation under refraction, the direction correction matrix and amplitude compensation factor are established for the E-vector. On this basis, the refracted E-vector is introduced into the heading measurement model, which can reduce modeling errors induced by refraction, thereby producing a better prediction of heading information. In addition, a dual-filter algorithm is constructed to handle fusion models with different characteristics, improving the reliability and efficiency of heading estimation. Numerical simulation is conducted to confirm the feasibility of the proposed method, while ocean navigation experiments on UUV are carried out to evaluate its accuracy.

Thin-walled parts are characterized by weak rigidity and are very prone to deformation during machining, which directly affects the machining accuracy and performance of the parts, so controlling the machining deformation of thin-walled parts is an urgent process problem to be solved. To address such problems, this paper proposes a flatness control machining method based on ice holding of workpieces. The method uses frozen suction cups to solidify liquid water to achieve stress-free clamping of workpieces. We conducted a low-temperature tensile test to investigate the low-temperature mechanical properties of the material. Comparative tests of ice-free and low-temperature ice-fixed milling were conducted to compare and analyze the changes of flatness and milling force in the two working conditions, to investigate the influence of machining parameters on the flatness of thin-walled parts, and to reveal the mechanism of low-temperature milling of ice-fixed workpieces. In addition, the low-temperature milling performance of titanium/aluminum alloy based on ice-fixation was compared. The results show that TC4 has good plasticity, high flexural strength ratio and strong resistance to deformation at low temperatures. Compared with no ice-holding, ice-holding machining effectively improves the flatness of the workpiece, and the order of the machining parameters affecting flatness is: feed rate > milling depth > spindle speed. The milling forces under ice-holding conditions were all greater than those without ice holding. The stiffness and hardness, resistance to damage and deformation of titanium alloy at low temperature are greater than those of aluminum alloy. This method provides a new method for high-precision machining of thin-walled parts.

The cable-driven redundant manipulator (CDRM) has a light slender and highly dexterous body, which is especially suitable for fine manipulation in confined spaces or unstructured environments. However, for dexterous manipulation and physical interaction with its surrounding environment, the force/torque sensor is indispensable, which will significantly increase the mass, dimension, and cost of the whole CDRM system. In this article, a force-sensing algorithm and compliance control framework for CDRM without a six-axis force/torque sensor are proposed. First, we design a modular cable-driven manipulator with two-level internal sensors, i.e., joint encoders and cable tension sensors. The multispace kinetic model is derived to establish the mapping between motor, cable, joint, and end-effector states. At the same time, we build a recursive dynamics model that takes the cables’ tension, cable–hole friction, links’ gravity, and end-effector forces into account. Then, the indirect force-sensing algorithm of the end-effector is proposed by combining the kinematic and dynamic equations and internal sensor information. Furthermore, a compliance controller based on indirect force sensing is designed. Finally, typical experiments are carried out based on the CDRM. Experiment results indicate that the force-sensing accuracy exceeds 95%, whereas the compliance controller demonstrates outstanding compliant behavior in human–robot interaction tasks.

Underwater autonomous navigation has long been a challenging problem due to the scarcity of information sources. The polarization navigation offers a feasible solution to this problem. The existing polarization navigation schemes, however, require that the horizontal attitude is known, which can only be used for 2-D orientation. To address the limitation, a 3-D attitude determination strategy is developed in this article by exploiting the underwater downwelling radiance fields (light intensity and polarization). In particular, the horizontal attitude information contained in the Snell's window, a unique underwater optical phenomenon induced by refraction, is extracted via an improved edge recognition method. On this basis, underwater polarization is exploited to calculate solar position for orientation. By this means, the 3-D attitude is acquired independently using underwater downwelling radiance fields. The effectiveness of the proposed strategy is validated via experiments in both the water tank and open sea environments.

The cable-driven redundant manipulator (CDRM) possesses remarkable flexibility and holds substantial potential for application in constrained environments. To ensure both the smooth movement of the end-effector during delicate operations and the safety of interactions with the surrounding environment, real-time sensing of forces acting on both the end and links is imperative. Current in situ sensor-based methods face limitations in their applicability to CDRMs due to size and load capacity constraints. Moreover, these methods fall short in measuring contact force and its location along the entire arm. In this article, we introduce an ex situ sensing approach for capturing the six-dimensional (6-D) force at the end and the contact force on the linkages of a CDRM. First, a multispace recursive dynamic model of the CDRM is established using the Newton–Euler method. This model establishes mapping relationships among cable tensions, joint torques, and operational forces at the end-effector. Then, a simplified dynamic model for the recursive subsystem is derived based on joint motion transmission relationships and recursive equations. This model decouples the dynamic equations and provides a versatile force-sensing model. It enables the realization of 6-D force/torque sensing at the end-effector, as well as the determination of the magnitude and location of external forces acting on the links. Finally, compliant controllers are designed based on different external force-sensing methods to cater to diverse operational requirements. Experimental validation of the proposed methods is conducted on a CDRM prototype. The results demonstrate that the accuracy of end-effector force sensing exceeds 95%, torque sensing surpasses 90%, and the positioning error of the link's contact force sensing is less than 20 mm. Furthermore, the compliance controllers exhibit excellent smoothness in tasks involving human–robot interaction.

As the main source of disturbances in gimbal servo systems, the dynamic rotor imbalance can induce a torque that deteriorates system performance. The imbalance disturbance is quasi-periodic with unknown frequency, phase and amplitude, and is also submerged in stochastic noises with unknown characteristics. In this brief, the online joint identification and estimation of the imbalance in gimbal servo systems is investigated. An improved online expectation-maximization (EM) framework is utilized, which consists of a particle filter to track the posterior distribution of the imbalance, variational Bayesian adaptive Kalman filters to deal with the uncertain noise statistics, and a gradient-based solver for frequency identification. Furthermore, the stochastic approximation technique is employed to facilitate online implementation. Experimental results demonstrate the effectiveness of the proposed scheme in estimating the imbalance disturbance, identifying the unknown frequency, and adapting to uncertain noise statistics.

This paper is concerned with the contact force estimation problem of robot manipulators based on imperfect dynamic models of the manipulator and the contact force. To handle the imperfect dynamic information of the manipulator, a hybrid model, consisting of the nominal model and the residual dynamics, is established for the manipulator, and the Gaussian process regression (GPR) technique is employed to learn the mean and covariance of the residual dynamics. On this basis, a virtual measurement equation is established for contact force estimation and a Gaussian process adaptive disturbance Kalman filter (GPADKF) is developed where the variational Bayes technique is employed to achieve online identification of the noise statistics in the force dynamics. The GPADKF is capable of decoupling the contact force from residual dynamics and system noises, thereby reducing the dependence on accurate dynamic models of the manipulator and the contact force. Simulation and experimental results demonstrate that the proposed scheme outperforms the state-of-art methods. Note to Practitioners—Contact force estimation for robot manipulators can be achieved by fusing the dynamic models of the manipulator and the contact force. When both models are imprecise, the traditional inverse dynamics-based and disturbance Kalman filter-based approaches can no longer provide accurate force estimates. To handle this challenge, a computationally efficient hybrid dynamic model is established for the manipulator, which consists of the nominal model and a residual dynamics compensation term learned from offline data via the GPR. On this basis, an adaptive disturbance Kalman filter is constructed by using the variational Bayes technique to deal with the inaccurate noise covariance matrix in the force dynamic model. Compared with the existing approaches, the force estimate obtained via the proposed scheme is more accurate and reliable, as refined noise covariance matrices (provided by both the GPR and the variational Bayes procedure) have been adopted in the Kalman gain calculation. The proposed GPADKF method is the extension of the composite disturbance filtering (CDF) framework. With the proposed scheme, the dependency on the perfect dynamic models in contact force estimation can be significantly reduced, and this makes our approach especially suitable for contact force estimation problems under unfamiliar and complicated environments.

Dynamic rotor imbalance is widely identified as the primary source of disturbance in gimbal servo systems and a major factor in the deterioration of their velocity tracking performance. The imbalance disturbance is often not directly measurable, submerged in noise, and with unknown frequency, which makes the estimation of such disturbances a particularly challenging topic. In order to mitigate the effects of the unknown imbalance, this paper investigates the disturbance identification problem, which includes simultaneous identification and estimation of the disturbance. Exploiting the features of the expectation-maximization (EM) framework, the disturbance identification problem is separated into the E-step (state estimation) and the M-step (model identification). A novel disturbance identification observer, where the E-step and the M-step are solved iteratively to simultaneously update the value and internal parameter of the disturbance online is proposed. In contrast to existing work using EM for identification of practical systems, the proposed scheme can be implemented online via stochastic approximation. In addition, a discrete-time anti-disturbance sliding mode controller based on the disturbance estimation is designed. Simulation and experimental results verify the effectiveness of the proposed method.

This article investigates the composite disturbance filtering (CDF) problem for a class of nonlinear stochastic systems subject to composite disturbances. The concerned disturbances include both the unknown deterministic type and the non-Gaussian stochastic type. In order to obtain the optimal state estimation under the influence of unknown input and non-Gaussian noise, a nonlinear CDF method is developed by resorting to the maximum correntropy criterion (MCC). Faced with the nonlinearity of system model as well as a conditionally linear substructure with respect to unknown input, the marginalized unscented transformation is exploited for computation-efficient statistics propagation, and then the statistical linearization is performed to provide a regression model for simultaneous state and unknown input estimation (SSUIE) under the MCC. The proposed filtering algorithm is demonstrated via a numerical example, and further applied to an integrated navigation system (INS). Simulation results confirm that our method has enhanced disturbance rejection ability and improved estimation accuracy in complex non-Gaussian environments.

The purpose of this work is to enhance KinasePhos, a machine learning-based kinase-specific phosphorylation site predictiontool. Experimentally verified kinase-specific phosphorylation data were collected from PhosphoSitePlus, UniProtKB, the GPS 5.0, and Phospho.ELM. In total, 41,421 experimentally verified kinase-specific phosphorylation sites were identified. A total of 1380 unique kinases were identified, including 753 with existing classification information from KinBase and the remaining 627 annotated by building a phylogenetic tree. Based on this kinase classification, a total of 771 predictive models were built at the individual, family, and group levels, using at least 15 experimentally verified substrate sites in positive training datasets. The improved models demonstrated their effectiveness compared with other prediction tools. For example, the prediction of sites phosphorylated by the protein kinase B, casein kinase 2, and protein kinase A families had accuracies of 94.5%, 92.5%, and 90.0%, respectively. The average prediction accuracy for all 771 models was 87.2%. For enhancing interpretability, the SHapley Additive exPlanations (SHAP) method was employed to assess feature importance. The web interface of KinasePhos 3.0 has been redesigned to provide comprehensive annotations of kinase-specific phosphorylation sites on multiple proteins. Additionally, considering the large scale of phosphoproteomic data, a downloadable prediction tool is available at https://awi.cuhk.edu.cn/KinasePhos/download.htmlor https://github.com/tom-209/KinasePhos-3.0-executable-file.

The purpose of this work is to enhance KinasePhos, a machine learning-based kinase-specific phosphorylation site predictiontool. Experimentally verified kinase-specific phosphorylation data were collected from PhosphoSitePlus, UniProtKB, the GPS 5.0, and Phospho.ELM. In total, 41,421 experimentally verified kinase-specific phosphorylation sites were identified. A total of 1380 unique kinases were identified, including 753 with existing classification information from KinBase and the remaining 627 annotated by building a phylogenetic tree. Based on this kinase classification, a total of 771 predictive models were built at the individual, family, and group levels, using at least 15 experimentally verified substrate sites in positive training datasets. The improved models demonstrated their effectiveness compared with other prediction tools. For example, the prediction of sites phosphorylated by the protein kinase B, casein kinase 2, and protein kinase A families had accuracies of 94.5%, 92.5%, and 90.0%, respectively. The average prediction accuracy for all 771 models was 87.2%. For enhancing interpretability, the SHapley Additive exPlanations (SHAP) method was employed to assess feature importance. The web interface of KinasePhos 3.0 has been redesigned to provide comprehensive annotations of kinase-specific phosphorylation sites on multiple proteins. Additionally, considering the large scale of phosphoproteomic data, a downloadable prediction tool is available at https://awi.cuhk.edu.cn/KinasePhos/download.htmlor https://github.com/tom-209/KinasePhos-3.0-executable-file.

This article is concerned with high-precision and fast velocity-tracking control for the harmonic drive-based gimbal system of control moment gyro (CMG). As a major challenge in the controller design, the simultaneous presence of multiple matched and mismatched disturbances should be carefully addressed. Specifically, the rotor imbalance inherent in the CMG will cause vibration of the output torque that significantly degrades the tracking performance. On the other hand, the integration of harmonic drives will lead to nonlinear torsional stiffness and introduce additional frictions into the control system. Furthermore, practical tasks such as fast attitude maneuver have posed higher requirement on the rapidity of the antidisturbance control schemes. To overcome the aforementioned difficulties, we propose a finite-time refined antidisturbance velocity tracking controller, which consists of a finite-time harmonic disturbance observer to compensate for the imbalance vibration, an adaptive finite-time extended state observer to handle the remaining disturbances with unknown rate of change, and a finite-time controller to ensure fast convergence of the tracking error. Both numerical simulation and hardware-in-the-loop experimental results show that the proposed scheme outperforms the existing methods in terms of accuracy and rapidity.

In this paper, with the aim of approximate the ball convergence of nonlinear equation, we study the local properties of a class of sixth-order Jarratt-type iterative methods in Banach spaces. By utilizing the first-order Fréchet derivative, we establish the local convergence. At last, the nonlinear equations related to the gas-state equation, the continuous stirred tank reactor (CSTR) and the problem of azeotropic point of a binary solutions are solved using this iterative method, and the applicability of the theoretical results is proved.

A composite anti-disturbance predictive control strategy employing a Multi-dimensional Taylor Network (MTN) is presented for unmanned systems subject to time-delay and multi-source disturbances. First, the multi-source disturbances are addressed according to their specific characteristics as follows: (A) an MTN data-driven model, which is used for uncertainty description, is designed accompanied with the mechanism model to represent the unmanned systems; (B) an adaptive MTN filter is used to remove the influence of the internal disturbance; (C) an MTN disturbance observer is constructed to estimate and compensate for the influence of the external disturbance; (D) the Extended Kalman Filter (EKF) algorithm is utilized as the learning mechanism for MTNs. Second, to address the time-delay effect, a recursive τ−step-ahead MTN predictive model is designed utilizing recursive technology, aiming to mitigate the impact of time-delay, and the EKF algorithm is employed as its learning mechanism. Then, the MTN predictive control law is designed based on the quadratic performance index. By implementing the proposed composite controller to unmanned systems, simultaneous feedforward compensation and feedback suppression to the multi-source disturbances are conducted. Finally, the convergence of the MTN and the stability of the closed-loop system are established utilizing the Lyapunov theorem. Two exemplary applications of unmanned systems involving unmanned vehicle and rigid spacecraft are presented to validate the effectiveness of the proposed approach.

Tuberculosis (TB) is a severe disease caused by Mycobacterium tuberculosisthat poses a significant threat to human health. The emergence of drug-resistant strains has made the global fight against TB even more challenging. Antituberculosis peptides (ATPs) have shown promising results as a potential treatment for TB. However, conventional wet lab-based approaches to ATP discovery are time-consuming and costly and often fail to discover peptides with desired properties. To address these challenges, we propose a novel machine learning-based framework called ATPfinder that can significantly accelerate the discovery of ATP. Our approach integrates various efficient peptide descriptors and utilizes the deep forest algorithm to construct the model. This neural network-like cascading structure can effectively process and mine features without complex hyperparameter tuning. Our experimental results show that ATPfinder outperforms existing ATP prediction tools, achieving state-of-the-art performance with an accuracy of 89.3% and an MCC of 0.70. Moreover, our framework exhibits better robustness than baseline algorithms commonly used for other sequence analysis tasks. Additionally, the excellent interpretability of our model can assist researchers in understanding the critical features of ATP. Finally, we developed a downloadable desktop application to simplify the use of our framework for researchers. Therefore, ATPfinder can facilitate the discovery of peptide drugs and provide potential solutions for TB treatment. Our framework is freely available at https://github.com/lantianyao/ATPfinder/(data sets and code) and https://awi.cuhk.edu.cn/dbAMP/ATPfinder.html(software).

In this paper, we propose a novel anti-disturbance attitude control law for combined spacecraft with an improved closed-loop control allocation scheme. More specifically, a saturated approach is adopted to guarantee the global asymptotic stability under control input saturation. To enhance the robustness of the system, a nonlinear disturbance observer is constructed to compensate the disturbances caused by inertial parameter uncertainty and unmodeled dynamics. Next, the quadratic programming algorithm is used to obtain an optimal open-loop control allocation scheme, where both energy consumption and actuator saturation have been considered in the allocation of the virtual control command. Then, a modified closed-loop control allocation scheme is proposed to reduce the allocation error under the actuator uncertainty. Finally, stability analysis of the closed-loop system with the proposed allocation scheme is provided. Simulation results confirm the effectiveness of the proposed control scheme.

Proton exchange membrane fuel cells (PEMFCs) are an emerging candidate for clean power generation, with applications in smart grids, vehicles, and unmanned aerial vehicles. Accurate modeling of the power characteristics of PEMFC is required for system management purposes, where an electro-chemical model with unknown parameters is often adopted. Estimation of the unknown parameters is challenging as operating condition shifts and system degradation will lead to variation in PEMFC characteristics, particularly in the presence of system nonlinearities and noise. In this paper, a novel online parameter estimation approach for PEMFC is proposed based on the marginalized particle filtering approach. By introducing a filter derivation based on Bayesian inference and estimating the linear and nonlinear parameters separately, the marginalized approach has reduced computation cost compared with conventional particle filters. Estimation accuracy of the proposed approach is validated by experimental and simulation results, where superior accuracy compared with extended Kalman filter is obtained. Furthermore, a self-designed hydrogen quadrotor was flight-tested and energy management studies were conducted to assess the performance of the proposed estimator in hydrogen unmanned aerial vehicle applications using real flight data.

The particle filtering algorithm was introduced in the 1990s as a numerical solution to the Bayesian estimation problem for nonlinear and non-Gaussian systems and has been successfully applied in various fields including physics, economics, engineering, etc. As is widely recognized, the particle filter has broad application prospects in networked systems, but network-induced phenomena and limited computing resources have led to new challenges to the design and implementation of particle filtering algorithms. In this survey paper, we aim to review the particle filtering method and its applications in networked systems. We first provide an overview of the particle filtering methods as well as networked systems, and then investigate the recent progress in the design of particle filter for networked systems. Our main focus is on the state estimation problems in this survey, but other aspects of particle filtering approaches are also highlighted.

Underwater navigation system is an indispensable part for autonomous underwater vehicles. Due to the indiscernibility of satellite signal, however, the underwater navigation problem is quite challenging, and a satellite-free navigation scheme should be looked for. Polarization navigation, inspired by insects’ capability of autonomous homing and foraging, is an alternative solution to satellite navigation with great application potential. Underwater polarization provides an indirect sun compass to animals for orientation determination. However, it is difficult to apply terrestrial solar-tracking methodologies in underwater situations due to the refraction of polarized skylight at the air–water interface. To resolve this issue, an underwater solar-tracking algorithm is developed based on the underwater refraction-polarization pattern inside the Snell's window. By employing Snell's law and Fresnel refraction formula to decouple the refractive ray bending and polarization deflection, the celestial polarization pattern is obtained based on underwater measurement. To further improve the accuracy, the degree of polarization is employed as a weight factor for E-vector. A long-lasting underwater experiment was conducted to validate the effectiveness of the proposed approach, and the results showed the root-mean-square errors of solar zenith and azimuth employing this algorithm were 0.3° and 1.3°, respectively. Our experimental results show that the refraction-polarization pattern inside the Snell's window exhibits immense potential to improve the solar-tracking accuracy for underwater navigation.

Bioinspired polarized skylight navigation, which can be used in unfamiliar territories, is an important alternative autonomous navigation technique in the absence of Global Navigation Satellite System (GNSS). However, the polarization pattern in night environment with noise effects and model uncertainties is a less explored area. Although several decades have passed since the first publication about the polarization of the moonlit night sky, the usefulness of nocturnal polarization navigation is only sporadic in previous researches. This study demonstrates that the nocturnal polarized light is capable of providing accurate and stable navigation information in dim light outdoor environment. Based on the statistical characteristics of Angle of Polarization (AoP) error, a probability density estimation method is proposed for heading determination. To illustrate the application potentials, the simulation and outdoor experiments are performed. Resultingly, the proposed method robustly models the distribution of AoP error and gives accurate heading estimation evaluated by Standard Deviation (STD) which is 0.32° in a clear night sky and 0.47° in a cloudy night sky.