In coal seam drilling, wall friction along the borehole introduces complex, spatially distributed disturbances that degrade the dynamic performance of the drill string. Due to its slender and flexible structure, the drill string is highly sensitive to such multi-point excitations, leading to frequent velocity fluctuations and reduced tracking accuracy. To address this, a robust H-infinity control strategy is proposed. A disturbance weighting function is designed to give the controller notch-filter characteristics, enabling targeted suppression of resonance-induced vibrations. Relying only on inlet measurements, the controller ensures accurate tracking of the reference feeding speed while effectively mitigating wall friction effects. Simulation results show significant improvements in steady-state accuracy, disturbance rejection, and robustness compared to conventional methods, confirming the effectiveness of the proposed approach.

A multibranch adaptive graph convolutional network is proposed for human action recognition by combining graph convolutional networks (GCNs), adaptive learning, and multibranch feature extraction. Through the adaptive graph convolution module, this method can adaptively change parameters during the training process, thereby enhancing the flexibility of the model. Furthermore, the integration of shallow-level features (skeleton joints), with deep-level features including skeleton information, motion information, and motion difference information allows our model to capture both spatial and temporal dynamics of human actions, leading to a more comprehensive representation of human action features. The introduction of the spatio-temporal attention mechanism enables our model to focus on key frames and skeleton joints. The attitude correction module makes the input data to the network more reasonable and reduces the interference of noise. The inclusion of the adaptive mechanism makes the network no longer limited to the inherent physical connections, and the flexibility of the network is enhanced. The addition of second-order features makes the features of the skeletal data fully exploited. This attention mechanism enhances the discriminative ability of the model and improves its ability to recognize subtle variations and important cues in human actions. Through experiments on benchmark datasets, NTU-RGB-D and Kinetics-400, our method achieves significant improvements in action recognition performance compared with existing approaches. On the Kinetics-400 dataset, we achieved 36.5% and 59.6% recognition rates under the Top-1 and Top-5 evaluation metrics, respectively, which is an improvement of about 1% compared with the state-of-the-art method. On the NTU-RGB-D dataset, we achieved 95.8% and 89.4% recognition rates under the X-view and X-subject modes, respectively, with excellent results. These results validate the effectiveness of the multi-branch adaptive graph convolutional network for human action recognition tasks.

It is inevitable to encounter through different formations in the drilling process for deep exploration, and the penetration resistance coefficient (PRC) is an uncertain parameter related to lithology. In this article, a parameter-estimation-based gain-scheduling controller is developed to eliminate undesired system performance deterioration due to the uncertain parameter. First, a drill-string axial finite element model with the uncertain PRC is established, and a control-oriented low-order model is derived via mode selection. A gain-scheduling controller is computed based on the quadratic stability condition of the closed-loop system, which can cope with the system's parameter uncertainty using variable low-frequency gain. An adaptive observer is designed to estimate the unmeasurable scheduling variable. Field data from a geothermal drilling well is obtained to validate our model. According to this drilling well scenario, both numerical and experimental results illustrate the effectiveness of our method. The explicit relationship between the controller gain and the uncertain parameter is presented. It is found that the performance of the closed-loop system is more sensitive to the controller gain when drilling in soft formations, requiring more attention in such scenarios.

This study presents a two-dimensional (2-D) repetitive control method to address the issues of periodic tracking and disturbance suppression in uncertain Takagi–Sugeno systems. The disturbance and uncertainty are treated as an equivalent-input-disturbance (EID). However, the conventional EID estimators typically suppress the EID through high gain. Meanwhile, the low-pass filter associated with EID causes a certain degree of phase lag. A proportional–integral (PI) filter is integrated with an EID estimator to develop a PI-EID structure to improve the estimation accuracy. Based on the self-learning mechanism of repetitive control, the 2-D repetitive controller is used to achieve a high level of tracking. Unlike the conventional nonlinear repetitive control methods, the state observer and the PI-EID estimator are membership function dependent. The gains of both controllers switch in line with the signs of the time derivative of the normalized premise variables, and this framework takes full account of the information of the nonlinear membership functions. The controller design procedures and the stability conditions are detailedly presented. Finally, a rotation speed control experiment is conducted to validate the developed PI-EID method.

An original attention-based indoor scene recognition model combining local and global features is proposed. Multi-strategy data augmentation using several different functions and intensities can improve the classification performance. Then, local features are extracted using a convolutional layer and a single self-attention, thus solving the problem of large intra-class variance. The multi-attention mechanism is used to fuse the local feature information extracted from different foci to obtain a more complete global feature representation. The multi-head attention mechanism allows the network to extract features in parallel in different directions of attention, which helps the network to better capture global information, improves the network's ability to understand and represent the input data, and solves the problem of high inter-class similarity. Finally, the extracted features are fed into the classifier to complete the classification of indoor scene images. Experiments are conducted on four data sets (IndoorCVPR09, SUN397, 15-Scenes and self-built small sample scientific indoor scene dataset), yield excellent results. The results show that the developed algorithm effectively solves the two problems of high intra-class diversity and high inter-class similarity. As a result, the model has achieved competitive results. Preliminary application experiments are developed in our HRI system, indicating that the proposed indoor scene recognition model can be applied to the complete environmental perception in HRI.

A coupled multimodal emotional feature analysis (CMEFA) method based on broad-deep fusion networks, which divide multimodal emotion recognition into two layers, is proposed. First, facial emotional features and gesture emotional features are extracted using the broad and deep learning fusion network (BDFN). Considering that the bi-modal emotion is not completely independent of each other, canonical correlation analysis (CCA) is used to analyze and extract the correlation between the emotion features, and a coupling network is established for emotion recognition of the extracted bi-modal features. Both simulation and application experiments are completed. According to the simulation experiments completed on the bimodal face and body gesture database (FABO), the recognition rate of the proposed method has increased by 1.15% compared to that of the support vector machine recursive feature elimination (SVMRFE) (without considering the unbalanced contribution of features). Moreover, by using the proposed method, the multimodal recognition rate is 21.22%, 2.65%, 1.61%, 1.54%, and 0.20% higher than those of the fuzzy deep neural network with sparse autoencoder (FDNNSA), ResNet-101 + GFK, C3D + MCB + DBN, the hierarchical classification fusion strategy (HCFS), and cross-channel convolutional neural network (CCCNN), respectively. In addition, preliminary application experiments are carried out on our developed emotional social robot system, where emotional robot recognizes the emotions of eight volunteers based on their facial expressions and body gestures.

Convolutional feature-based broad learning with long short-term memory (CBLSTM) is proposed to recognize multidimensional facial emotions in human–robot interaction. The CBLSTM model consists of convolution and pooling layers, broad learning (BL), and long- and short-term memory network. It aims to obtain the depth, width, and time scale information of facial emotion through three parts of the model, so as to realize multidimensional facial emotion recognition. CBLSTM adopts the structure of BL after processing was done at the convolution and pooling layer to replace the original random mapping method and extract features with more representation ability, which significantly reduces the computational time of the facial emotion recognition network. Moreover, we adopted incremental learning, which can quickly reconstruct the model without a complete retraining process. Experiments on three databases are developed, including CK+, MMI, and SFEW2.0 databases. The experimental results show that the proposed CBLSTM model using multidimensional information produces higher recognition accuracy than that without time scale information. It is 1.30% higher on the CK+ database and 1.06% higher on the MMI database. The computation time is 9.065 s, which is significantly shorter than the time reported for the convolutional neural network (CNN). In addition, the proposed method obtains improvement compared to the state-of-the-art methods. It improves the recognition rate by 3.97%, 1.77%, and 0.17% compared to that of CNN-SIPS, HOG-TOP, and CMACNN in the CK+ database, 5.17%, 5.14%, and 3.56% compared to TLMOS, ALAW, and DAUGN in the MMI database, and 7.08% and 2.98% compared to CNNVA and QCNN in the SFEW2.0 database.

The two-stage fuzzy fusion based-convolution neural network is proposed for dynamic emotion recognition by using both facial expression and speech modalities, which not only can extract discriminative emotion features which contain spatio-temporal information, but also can effectively fuse facial expression and speech modalities. Moreover, the proposal is able to handle situations where the contributions of each modality data to emotion recognition are very imbalanced. The local binary patterns coming from three orthogonal planes and spectrogram are considered first to extract low-level dynamic emotion, so that the spatio-temporal information of these modalities can be obtained. To reveal more discriminative features, two deep convolution neural networks are constructed to extract high-level emotion semantic features. Moreover, the two stage fuzzy fusion strategy is developed by integrating canonical correlation analysis and fuzzy broad learning system, so as to take into account the correlation and difference between different modal features, as well as handle the ambiguity of emotional state information. The experimental results obtained on benchmark databases show that the accuracies of the proposed method are higher than those of existing methods (such as the hybrid deep model, and the rule-based and machine learning method) on SAVEE, eNTERFACE’05, and AFEW databases.

Sintering is the preproduction process of ironmaking, whose products are the basis of ironmaking. How to improve the operating performance of the iron ore sintering process has always been a problem that operators are committed to solve. An operating performance improvement method based on prediction and grade assessment is presented in this article. First, considering the data distribution characteristics of the process, a performance index prediction model based on the Gaussian process regression is built, in which the mutual information analysis method is used to select the inputs of the performance index prediction model. Then, the operating performance grade is assessed by a threshold division method. Next, the operating performance grade guides the control of the burn-through point to improve the operating performance. Finally, experimental verification is performed based on the actual running data. The results show that the proposed method has high prediction accuracy, and it is also significant in improving the operating performance. Therefore, this approach provides an effective solution to predict and improve operating performance.

A key energy consumption in steel metallurgy comes from an iron ore sintering process. Enhancing carbon utilization in this process is important for green manufacturing and energy saving and its prerequisite is a time-series prediction of carbon efficiency. The existing carbon efficiency models usually have a complex structure, leading to a time-consuming training process. In addition, a complete retraining process will be encountered if the models are inaccurate or data change. Analyzing the complex characteristics of the sintering process, we develop an original prediction framework, that is, a weighted kernel-based fuzzy C-means (WKFCM)-based broad learning model (BLM), to achieve fast and effective carbon efficiency modeling. First, sintering parameters affecting carbon efficiency are determined, following the sintering process mechanism. Next, WKFCM clustering is first presented for the identification of multiple operating conditions to better reflect the system dynamics of this process. Then, the BLM is built under each operating condition. Finally, a nearest neighbor criterion is used to determine which BLM is invoked for the time-series prediction of carbon efficiency. Experimental results using actual run data exhibit that, compared with other prediction models, the developed model can more accurately and efficiently achieve the time-series prediction of carbon efficiency. Furthermore, the developed model can also be used for the efficient and effective modeling of other industrial processes due to its flexible structure.

Information granules can discover interpretable and meaningful relationships offering a full description for time series. This article presents an information granulation method with rectangular information granules and applies it to time-series similarity measurement. First, the fuzzy$c$-means clustering algorithm transforms the time series and its first-order difference time series to data clusters. With the maximum volume of rectangular information granules viewed as the criterion, the optimal rectangular information granules are formed using the data cluster by the principle of justifiable granularity and the gravitational search algorithm. The time-series similarity is measured by calculating the similarity between the upper and lower bounds of the optimal rectangular information granules built from the time series. Finally, an experiment is performed on a public dataset to verify the feasibility of the proposed method. The result shows that the rectangular information granulation method can capture the change characteristics of time series. The similarity measurement method can effectively evaluate the similarity of the time series belonging to different classes.

This article is concerned with the trajectory optimization problem of a drilling process, which is of significance to drilling efficiency and safety. Due to the difference between an actual trajectory and a planned trajectory, this problem is a multiobjective optimization problem (MOP) with parameter uncertainties. To solve the problem, this study devises a new approach named nondominated sorting genetic Algorithm II using outlier removal (OR-NSGA-II). First, the optimization problem with parameter uncertainties is formulated, including two objective functions: 1) a trajectory length and 2) an expected value of drill-string torque. Then, an outlier-removal mechanism is devised in the sorting process of NSGA-II to reduce the negative effects of parameter uncertainties. Next, the crowding distance calculation in NSGA-II is improved to ensure population diversity. Comparison of simulation results show that our method is effective on the MOP with parameter uncertainties.

The safe and efficient operation of geological drilling systems is critically dependent on proper process monitoring. A monitoring model based on probabilistic principal component analysis is presented to deeply exploit performance feature information hidden among the process data in this paper. First, the offline monitoring model is established with historical process data collected from different sources to reveal different characteristics, and the monitoring statistics as a benchmark are obtained. Then, actual operating data are introduced to the established model to realize online monitoring. Finally, the monitoring effect is discussed, and the causes of inefficient drilling are tracked. The experimental results indicate that the proposed method can effectively monitor the operating performance of the drilling process.

The aim of this paper is to implement a robust weight on bit control design in an unified experimental system. A laboratory drilling rig is used to simulate the real world bit-rock interaction. Via communication between programmable logic controller (PLC) and OLE for Process Control (OPC) server, the laboratory drilling rig can exchange data with a finite element drill-string model and a hoisting system model which are realized in MATLAB/Simulink. Thus, an unified experimental system is established to reproduce the actual drilling process. The penetration resistance coefficient of the experiment rock sample is tested to obtain the bit-rock interaction model for controller design. Considering the high-order mode of the slender drill-string, a reduced order model with multiplicative weighting function is derived. Based on the developed model, a robust integrated controller is obtained by connecting a PI controller and a dynamic output feedback controller. Experiment results are presented to show the effectiveness of our method and the established system.

The weight-adapted convolution neural network (WACNN) is proposed to extract discriminative expression representations for recognizing facial expression. It aims to make good use of the convolution neural network's (CNN's) potential performance in avoiding local optima and speeding up convergence by the hybrid genetic algorithm (HGA) with optimal initial population, in such a way that it realizes deep and global emotion understanding in human-robot interaction. Moreover, the idea of novelty search is introduced to solve the deception problem in the HGA, which can expend the search space to help genetic algorithm jump out of local optimum and optimize large-scale parameters. In the proposal, the facial expression image preprocessing is conducted first, then the low-level expression features are extracted by using a principal component analysis. Finally, the high-level expression semantic features are extracted and recognized by WACNN which is optimized by HGA. In order to evaluate the effectiveness of WACNN, experiments on JAFFE, CK+, and static facial expressions in the wild 2.0 databases are carried out by using k -fold cross validation, and experimental results show the recognition accuracies of the proposal are superior to that of the state-of-the-art, such as local directional ternary pattern and weighted mixture deep neural network (DNN), which aim to extract discriminative and are the DNN-based methods. Moreover, recognition accuracies of the proposal are also higher than the deep CNN without HGA, which indicates that the proposal has better global optimization ability. Meanwhile, preliminary application experiments are also carried out by using the proposed algorithm on the emotional social robot system, where nine volunteers and two-wheeled robots experience the scenario of emotion understanding. Application results indicate that the wheeled robots can recognize basic expressions, such as happy, surprise, and so on.

A fuzzy deep neural network with sparse autoencoder (FDNNSA) is proposed for intention understanding based on human emotions and identification information (i.e., age, gender, and region), in which the fuzzy C-means (FCM) is used to cluster the input data, and deep neural network with sparse autoencoder (DNNSA) is designed for emotional intention understanding in human-robot interaction. It aims to make robots capable of recognizing human emotions and understanding related emotional intention, the FCM is suitable for gathering similar information so that the calculations of dimensionality of DNNSA will be reduced, and the sparse autoencoder of DNNSA can make the neuron of DNNSA sparse to reduce the complexity of the network in such a way human-robot interaction is running smoothly. To validate the proposal, simulation experiments based on benchmark databases such as facial expression database of CK+, and speech emotion corpus of CASIA were completed. The experimental results show that the proposal outperforms the baseline algorithms of Softmax regression (SR), DNNSA, FCM-based SR (FSR), Softplus, Gath Geva-based DNNSA (GDNNSA), and ensemble DNNSA (EDNNSA). Preliminary application experiments are performed in the development of emotional social robot system, where volunteers experience the scenario of “drinking at the bar”. The obtained results indicate that the proposed FDNNSA can promote robot understanding of emotional intention of human.

Sinter ore is the main raw material for ironmaking, and burn-through point (BTP) is one of the significant factors to measure the stability of the sintering process. In this article, through the feature extraction of time-series trend, a fuzzy control strategy is presented for the BTP. First, the Hurst exponent of the time series for the BTP is calculated by resorting to the rescaled range analysis method, by which the trend feature is analyzed. Then, by using the Mann-Kendall test, both global and local trend feature variable of the time series for the BTP are extracted and regarded as the inputs of the fuzzy controller. Next, a fuzzy controller for the BTP is designed to produce the control quantity of the strand velocity. Finally, based on a semiphysical simulation system and the raw data collected from an iron and steel plant, an experiment is carried out to demonstrate the effectiveness of the proposed control strategy.

Two-layer fuzzy support vector regression-Takagi-Sugeno (TLFSVR-TS) model is proposed for emotion understanding in human-robot interaction (HRI), where the real-time dynamic emotion is recognized according to facial expression, and emotional intention understanding is obtained mainly based on human emotions and identification information. It aims to make robots capable of recognizing and understanding human emotions, in such a way that make HRI run smoothly. TLFSVR-TS considers about the priori knowledge inferred from human personal preference to reduce the uncertainty of various people, and multiple support vector regression (SVR) corresponding to different genders/provinces/ages of human to guarantee the local learning ability. Preliminary application experiments are performed in the developing emotional social robot system, where 30 volunteers experience the scenario of “drinking in the bar.” Results show that the proposal receives higher understanding accuracy than that of TLFSVR, kernel fuzzy c-means clustering is fused with SVR, and SVR.

This paper develops a quick and effective continuous state feedback control strategy based on intelligent optimization for a planar three-link passive–active–active (PAA) underactuated system with first-order nonholonomic characteristic. We analyze the integral characteristic of the system and find the passive link will be stabilized at different angle when the two active links of the system are controlled to different angles. Therefore, we design a set of continuous state feedback controllers based on the control targets of the two active links, and then we optimize the target angles of all links and the design parameters of the above controllers by employing intelligent optimization algorithm. In this way, the system can be stabilized at the target point by using these controllers with optimized parameters. Finally, the simulation results demonstrate that the continuous control strategy based on intelligent optimization may make the planar PAA underactuated system be stabilized at any target point from any initial point.

A three-layer weighted fuzzy support vector regression (TLWFSVR) model is proposed for understanding human intention, and it is based on the emotion-identification information in human-robot interaction. The TLWFSVR model consists of three layers, including adjusted weighted kernel fuzzy c-means for data clustering, fuzzy support vector regressions (FSVR) for information understanding, and weighted fusion for intention understanding. It aims to guarantee the quick convergence and satisfactory performance of the local FSVR via adjusting the weights of each feature in each cluster, in such a way that importance of different emotion-identification information is represented. Moreover, smooth human-oriented interaction can be obtained by endowing robot with human intention understanding capability. Experimental results show that the proposed TLWFSVR model obtains higher intention understanding accuracy and less computational time than that of two-layer fuzzy support vector regression, support vector regression, and back propagation neural network (BPNN), respectively. Additionally, the preliminary application experiments are performed in the developing human-robot interaction system, called emotional social robot system, where 12 volunteers and 2 mobile robots experience a scenario of “drinking at a bar.” Application results indicate that the bartender robot is able to understand customers' order intentions.

A two-layer fuzzy kernel regression (TLFKR) model is proposed for understanding human emotional intention in human-robot interaction, where TLFKR model consists of two layers, including fuzzy c-means (FCM) with kernel ridge regression (Kernel 1) for information analysis layer, fuzzy support vector regressions (FSVR) (Kernel 2) for intention understanding layer. TLFKR model represents the weight impact for each emotional information and aims to improve smooth human-robot interaction by endowing robot with human emotional intention understanding capability. Experimental Results show that the proposal obtains an intention understanding accuracy of 65.67%/68.33%/80.67% with the clusters number c=2/3/6 (according to different genders/ages/nationalities), which are 7.34%/7.18%/8.67% and 18.67%/21.33%/33.67% higher than that of TLFSVR and SVR, respectively. Additionally, preliminary application experiments are performed in the developing emotional social robot system, where two mobile robots and volunteers experience a scenario of “drinking at a bar”, and social robots are able to express basic emotions and understand human order intention.

To ensure that people can supplement enough water at the right time, an initiative service method based on intention understanding for drinking service robot is proposed. Firstly, individual factors and environmental factors which are associated with personal intentions are collected. Secondly, users' intention and desire degree for intention are obtained by an improved Fuzzy-Context-specific Intention Inference method. Then the relationship between personal intention and human demands is established by an initiative service model. Finally, robots provide drinking water service initiatively according to humans' demands. Drinking service experiments are performed in a laboratory scenario using a humans-robots interaction system. The experimental results show that the drinking time of actual situation lags behind the initiative service time 25 minutes on average, and the initiative service renders drinking service 2 times more than actual situation on average, which demonstrates the feasibility of the proposal. In prospect, the initiative service method could be applied to many occasions in our daily life, e.g., family service, caring for the elderly, and medical rehabilitation.

Adaptive dynamic programming (ADP) is a prevalent way to solve the coupled Hamilton-Jacobi-Bellman (HJB) equations of the optimal consensus control for multi-agent systems (MAS). Neural networks (NNs) are normally used to approximate the value functions in ADP. However, NNs with manually designed features may influence the approximation ability. In this study, kernel-based methods which do not need to set the value function model structure in advance are adopted for value functions approximation. Moreover, to overcome the deficiency that most of the system dynamics are unknown, or the system is too complex to obtain the accurate dynamics. Local action value functions are defined, and kernel-based methods are used to approximate the local action value functions. Thus, an action dependent heuristic dynamic programming (ADHDP) approach using kernel-based local action value functions approximation is developed to achieve the optimal consensus control model-freely. The developed approach uses historical sample data to learn the system dynamics, and avoids the traditional system identification scheme. Simulation results are provided to demonstrate the effectiveness of the presented approach.