Search

BackgroundClinical notes contain contextualized information beyond structured data related to patients’ past and current health status. ObjectiveThis study aimed to design a multimodal deep learning approach to improve the evaluation precision of hospital outcomes for heart failure (HF) using admission clinical notes and easily collected tabular data. MethodsData for the development and validation of the multimodal model were retrospectively derived from 3 open-access US databases, including the Medical Information Mart for Intensive Care III v1.4 (MIMIC-III) and MIMIC-IV v1.0, collected from a teaching hospital from 2001 to 2019, and the eICU Collaborative Research Database v1.2, collected from 208 hospitals from 2014 to 2015. The study cohorts consisted of all patients with critical HF. The clinical notes, including chief complaint, history of present illness, physical examination, medical history, and admission medication, as well as clinical variables recorded in electronic health records, were analyzed. We developed a deep learning mortality prediction model for in-hospital patients, which underwent complete internal, prospective, and external evaluation. The Integrated Gradients and SHapley Additive exPlanations (SHAP) methods were used to analyze the importance of risk factors. ResultsThe study included 9989 (16.4%) patients in the development set, 2497 (14.1%) patients in the internal validation set, 1896 (18.3%) in the prospective validation set, and 7432 (15%) patients in the external validation set. The area under the receiver operating characteristic curve of the models was 0.838 (95% CI 0.827-0.851), 0.849 (95% CI 0.841-0.856), and 0.767 (95% CI 0.762-0.772), for the internal, prospective, and external validation sets, respectively. The area under the receiver operating characteristic curve of the multimodal model outperformed that of the unimodal models in all test sets, and tabular data contributed to higher discrimination. The medical history and physical examination were more useful than other factors in early assessments. ConclusionsThe multimodal deep learning model for combining admission notes and clinical tabular data showed promising efficacy as a potentially novel method in evaluating the risk of mortality in patients with HF, providing more accurate and timely decision support.

Optimal energy transmission dispatching of microgrid systems involves complicated transmission energy allocation and battery charging/discharging management and remains a difficult and challenging research problem subject to complex operation conditions and action constraints due to the randomness and volatility of new energy. Traditional microgrid transmission dispatching mainly considers the matching of the demand side and the supply side from a macro perspective, without considering the impact of line loss. Therefore, a Hierarchical Deep Q-network (HDQN) approach for microgrid energy dispatching is proposed to address this issue. The approach takes the power flow of each line and the battery charging/discharging behavior as decision variables to minimize the system operation cost. The proposed approach employs a two-layer agent optimization architecture for simultaneously processing the discrete and continuous variables, with one agent making upper layer decisions on the charging and discharging behavior of the batteries, and the other agent making lower layer decisions on the transmission energy allocation for the line. The experimental results indicate that the proposed approach achieves better performance than existing approaches.

With the widespread use of WiFi devices and the availability of channel state information (CSI), CSI-based device-free localization (DFL) has attracted lots of attention. Fingerprint-based localization methods are the primary solutions for DFL, but they are faced with the fingerprint similarity problem due to the complex environment and low bandwidth of the commercial WiFi. Meanwhile, fingerprints may change unpredictably due to multipath WiFi signal propagation in time-varying environments. To tackle these problems, this paper proposes an adaptive online learning DFL method, which adaptively updates the localization model to ensure long-term accuracy and adaptability. Specifically, the CSI signals of the target located at different reference points are first collected and transformed to discriminable fingerprints using the weights of Multilayer Online Sequence Extreme Learning Machine (ML-OSELM). After that, an online learning DFL model is built to adapt to the changes of the environment. Experimental results in a time-varying indoor environment validate the adaptability of the proposed method against environmental changes and show that our method can achieve 10% improvement over other methods.

Online purchasing is increasing because customers are shifting to digital wallets and digital money, as these services are provided by different microfinance and other commercial banking sectors, and different online brands are working in Pakistan to support environmental sustainability. The objective of this study was to demonstrate to what extent low distribution charges and low transit time is contributing to impulsive buying when customers can use digital money in Pakistan. The study was conducted using survey research. Importantly, 650 questionnaires were distributed to the respondents with a received response rate of 40%. The study found that digital money (e-wallet) is positively associated with impulsive buying. Moreover, the moderating role of distribution charges and low transit time has been significant in impulsive buying has been significant. This study concludes that low transit time and load distribution charges must be considered by online businesses and brands working in Pakistan to ensure productivity and capture a larger market share of impulsive buying in Pakistan. Also, the current study contributes a theoretical framework to the knowledge and literature related to impulsive buying. The scope of this study is limited to the online businesses and brands that are working to provide products and services to the Pakistani people with the help of digital money and digital transactions. Significantly, this study provides significant future directions that are important to consider for upcoming studies to focus on and contribute to effectively.

Wireless energy transfer technology (WET)-enabled mobile charging provides an innovative strategy for energy replenishment in wireless rechargeable sensor networks (WRSNs), where the mobile charger (MC) can charge the sensors sequentially by WET according to the mobile charging scheduling scheme. Although there have been fruitful studies, they usually assume that all sensors will be charged fully once scheduled or charged to a fixed percentage determined by a charging upper threshold, resulting in low charging performance as they cannot adjust the charging operation on each sensor adaptively according to the real-time charging demands. To tackle this challenge, we first formulate the mobile charging scheduling as a joint mobile charging sequence scheduling and charging upper threshold control problem (JSSTC), where the charging upper threshold of each sensor can adjust adaptively. Then, we propose a novel multi-discrete action space deep Q-network approach for JSSTC (MDDRL-JSSTC), where MC is regarded as an agent exploring the environment. The state information observed by MC at each time step is encoded to construct a high-dimensional vector. Furthermore, a two-dimensional action is mapped to the charging destination of MC and the corresponding charging upper threshold at the next time step, using bidirectional gated recurrent units (Bi-GRU). Finally, we conduct a series of experiments to verify the superior performance of the proposed approach in prolonging the lifetime compared with the state-of-the-art approaches.

Visual navigation based on deep reinforcement learning requires a large amount of interaction with the environment, and due to the reward sparsity, it requires a large amount of training time and computational resources. In this paper, we focus on sample efficiency and navigation performance and propose a framework for visual navigation based on multiple self-supervised auxiliary tasks. Specifically, we present an LSTM-based dynamics model and an attention-based image-reconstruction model as auxiliary tasks. These self-supervised auxiliary tasks enable agents to learn navigation strategies directly from the original high-dimensional images without relying on ResNet features by constructing latent representation learning. Experimental results show that without manually designed features and prior demonstrations, our method significantly improves the training efficiency and outperforms the baseline algorithms on the simulator and real-world image datasets.

Collaborative filtering (CF) is the most classical method for recommender system, but it is usually suffered from limited performance by the sparseness of user-to-item rating data. Recently, due to the powerful learning feature representation ability, deep learning components are used to leverage auxiliary information to assist recommendation. However, most existing models based on deep learning are incomplete so that merely extracting the item latent representation and ignoring the user parts. Besides, different data are not chosen from current models. This paper proposes a novel probability framework, named as joint matrix factorization (JMF). There are three components in JMF. Firstly, the modified multilayer crossing version of the factorization machine (MFM) is designed to extract the user latent factors based on user behavior information. Moreover, MFM is a general method which can be used to accomplish many tasks in terms of machine learning. Secondly, a modification of Long Short-Term Memory (LSTM), named as bidirectional LSTM (BLSTM), is used to extract the item latent factors of a document sequence from both front and back directions. Finally, we tightly integrate BLSTM and MFM into probabilistic matrix factorization (PMF) to form JMF. Compared with the classical matrix factorization and other integration models, JMF extracts document data as well as user behavioral data as item vectors and user vectors. Extensive experiments on five real-world datasets show the proposed model has better performance compared with the state-of-the-art recommendation methods.

In wireless rechargeable sensor networks (WRSNs), a novel approach to energy replenishment is offered by the utilization of mobile chargers (MCs), which charge nodes via wireless energy transfer technology. However, previous research on mobile charging schemes has commonly prioritized charging efficiency as a performance index, neglecting the importance of quality of sensing coverage (QSC). As the network scale increases, the MC’s charging power becomes unable to meet the energy needs of all nodes, leading to a decline in network QSC when nodes’ energy is depleted. To solve this problem, we study the problem of mobile charging sequence scheduling for optimal network QSC (MSSQ) and propose an improved quantum-behaved particle swarm optimization (IQPSO) algorithm. With the attraction of potential energy in quantum space, this algorithm will adaptively adjust the contraction expansion coefficient iteratively, leading to a global optimal solution for the mobile charging sequence. Extensive simulation results demonstrate the superiority of IQPSO over the widely used QPSO and Greedy algorithms in terms of network QSC, especially in large-scale networks.

The accurate identification of the shape of the blast furnace (BF) burden surface is a crucial factor in the fault diagnosis of the BF condition and guides the charge operation. Based on the BF 3D point cloud data of phased array radar, this paper proposes a 3D burden surface feature definition system. Based on expert experience, the feature parameters of the burden surface are extracted. The voxel feature was extracted based on improved BNVGG. The optimized PointCNN extracts the point cloud features. The features of the burden surface were defined from three perspectives: the surface shape, voxel, and point cloud. The research of the 2D burden line is extended to the 3D burden surface, and the assumption of the symmetry of the BF is eliminated. Finally, the accuracy of the burden surface classification under each feature was evaluated, and the effectiveness of each feature extraction algorithm was verified. The experimental results show that the shape feature defined based on expert experience affects the recognition of the burden surface. However, it is defined from the data perspective and cannot accurately identify a similar burden surface shape. Therefore, the voxel features and point cloud features of the burden surface were extracted, improving the identification accuracy.

The whale optimization algorithm(WOA) is a novel meta-heuristic evolutionary algorithm inspired by the behavior of whales predation. An important factor to the success of WOA is the balancing between exploration and exploitation. In the WOA, the distance control parameter $a$ is the main factor to find an appropriate balance between exploration and exploitation. In the standard WOA, the distance control parameter $a$ is optimized by linear control strategy (LCS), but the process of whales predation is not simply linear process. To address the issue, this paper proposed a nonlinear control strategy based on arcsine function (NCS-Arcsin) to optimize WOA. The NCS-Arcsin is applied to adjust distance control parameter $a$ . The NCS-Arcsin is considered to accurately describe the process of whales predation. Experiments on twelve well-known benchmark functions show the NCS-Arcsin can significantly improve the exploration and exploitation capabilities of WOA. In addition, the performance of proposed NCS-Arcsin is compared with LCS and other have been proposed NCS. The experimental results show that the optimization effect of NCS-Arcsin is stronger than that of LCS and other NCS. The NCSs based on the arcsine function is the best NCS, which can significantly improve the optimize performance of WOA.

The breakthrough of wireless energy transmission (WET) technology has greatly promoted the wireless rechargeable sensor networks (WRSNs). A promising method to overcome the energy constraint problem in WRSNs is mobile charging by employing a mobile charger to charge sensors via WET. Recently, more and more studies have been conducted for mobile charging scheduling under dynamic charging environments, ignoring the consideration of the joint charging sequence scheduling and charging ratio control (JSSRC) optimal design. This paper will propose a novel attention-shared multi-agent actor–critic-based deep reinforcement learning approach for JSSRC (AMADRL-JSSRC). In AMADRL-JSSRC, we employ two heterogeneous agents named charging sequence scheduler and charging ratio controller with an independent actor network and critic network. Meanwhile, we design the reward function for them, respectively, by considering the tour length and the number of dead sensors. The AMADRL-JSSRC trains decentralized policies in multi-agent environments, using a centralized computing critic network to share an attention mechanism, and it selects relevant policy information for each agent at every charging decision. Simulation results demonstrate that the proposed AMADRL-JSSRC can efficiently prolong the lifetime of the network and reduce the number of death sensors compared with the baseline algorithms.

There are significant regional differences in the development of China’s insurance industry. An important question is that such regional differences are expanding or shrinking? Based on Barrow’s economic growth convergence model, this paper uses the σ convergence model to analyze the differences in the development of China’s insurance industry and its trends. It draws on the statistics data from 1990 to 2020. The empirical results show that the convergence of China’s insurance development is not obvious before 2006, but it shows a significant convergence after 2006. And, there are some differences between the Eastern, the Central, and the Western. Furthermore, when considering the spatial correlation, the convergence of insurance development among provinces in China is more obvious. This shows that the flowing of capital, technology, and labor force between regions may be beneficial to the balanced development of insurance among the regions.

Energy storage is an important adjustment method to improve the economy and reliability of a power system. Due to the complexity of the coupling relationship of elements such as the power source, load, and energy storage in the microgrid, there are problems of insufficient performance in terms of economic operation and efficient dispatching. In view of this, this paper proposes an energy storage configuration optimization model based on reinforcement learning and battery state of health assessment. Firstly, a quantitative assessment of battery health life loss based on deep learning was performed. Secondly, on the basis of considering comprehensive energy complementarity, a two-layer optimal configuration model was designed to optimize the capacity configuration and dispatch operation. Finally, the feasibility of the proposed method in microgrid energy storage planning and operation was verified by experimentation. By integrating reinforcement learning and traditional optimization methods, the proposed method did not rely on the accurate prediction of the power supply and load and can make decisions based only on the real-time information of the microgrid. In this paper, the advantages and disadvantages of the proposed method and existing methods were analyzed, and the results show that the proposed method can effectively improve the performance of dynamic planning for energy storage in microgrids.

Click-through rate (CTR) prediction plays a predominant role in the online advertisements. CTR prediction is a problem of binary classification with imbalanced data. Many existing approaches for imbalance learning only focus on over-sampling and under-sampling, but these methods definitely ignore some vital information of the original data. In this paper, we first propose a weighted output extreme learning machine (WO-ELM) to learn the imbalanced data. A hierarchical extreme learning machine (H-C-ELM) is proposed based on the proposed WO-ELM and the weighted extreme learning machine (W-ELM). The H-C-ELM has two levels in its structure. In the first level, the WO-ELM and the W-ELM are trained on different combined fields of the CTR (each field has some attributes). The two extreme learning machines (ELMs) output their predicted scores of the corresponding combined fields of the CTR. The WO-ELM and the W-ELM have different predicted results on the same combined fields because of the difference of the two ELMs. Therefore, in the second level, another ELM is applied based on the outputs of the two ELMs in the first level and the actual outputs in order to improve the prediction accuracy. The experimental results demonstrate that the proposed H-C-ELM method has better performance for the binary classification with imbalanced data than the other related algorithms on CTR prediction, such as the WO-ELM, the W-ELM, and the stacked autoencoder-logistic regression.

Extreme learning machine (ELM) is a fast learning algorithm for the single-hidden layer feedforward neural networks. However, usually we cannot guarantee the stability of the ELM because the parameters of the ELM are generated randomly, such as its biases of the hidden layer and the connecting weights between the input layer and the hidden layer. Besides, it is hard for a single model to achieve high predicted accuracy on the dataset with low-quality data. In this paper, we first propose a modified residual ELM (R-ELM) to improve the ELM's learning performance. In R-ELM, the first ELM is trained by the original dataset and the m-th (m > 1) ELM will be trained by the residuals between the ground truths and the predicted results of the previous ensemble model (with m - 1 ELMs). R-ELM (with m ELMs) is built in the direction of the error reduction by calculating the m-th ELM's optimal weight which is determined by the loss function of the R-ELM. As a result, R-ELM can remember almost all information of the training set. However, this ability does not assure a similar performance on the testing dataset. In view of this problem, we add L2 regularization to the loss function of the R-ELM (RR-ELM) to avoid the overfitting problem of R-ELM. In RR-ELM, L2 regularization is employed to encourage each ELM to ignore the unnecessary information of the training set. In order to verify the effectiveness of the two proposed algorithms, the real data from the blast furnace are engaged to perform the experiments. Experimental results illustrate that the proposed RR-ELM and the R-ELM are more stable than the single ELM. These results also demonstrate that the two proposed methods are more accurate than the average outputs of a group of ELMs, the ELM, and the support vector regression.

Battery energy storage technology is an important part of the industrial parks to ensure the stable power supply, and its rough charging and discharging mode is difficult to meet the application requirements of energy saving, emission reduction, cost reduction, and efficiency increase. As a classic method of deep reinforcement learning, the deep Q-network is widely used to solve the problem of user-side battery energy storage charging and discharging. In some scenarios, its performance has reached the level of human expert. However, the updating of storage priority in experience memory often lags behind updating of Q-network parameters. In response to the need for lean management of battery charging and discharging, this paper proposes an improved deep Q-network to update the priority of sequence samples and the training performance of deep neural network, which reduces the cost of charging and discharging action and energy consumption in the park. The proposed method considers factors such as real-time electricity price, battery status, and time. The energy consumption state, charging and discharging behavior, reward function, and neural network structure are designed to meet the flexible scheduling of charging and discharging strategies, and can finally realize the optimization of battery energy storage benefits. The proposed method can solve the problem of priority update lag, and improve the utilization efficiency and learning performance of the experience pool samples. The paper selects electricity price data from the United States and some regions of China for simulation experiments. Experimental results show that compared with the traditional algorithm, the proposed approach can achieve better performance in both electricity price systems, thereby greatly reducing the cost of battery energy storage and providing a stronger guarantee for the safe and stable operation of battery energy storage systems in industrial parks.

Grey wolf optimizer (GWO) is a meta-heuristic algorithm inspired by the hierarchy of grey wolves (Canis lupus). Fireworks algorithm (FWA) is a nature-inspired optimization method mimicking the explosion process of fireworks for optimization problems. Both of them have a strong optimal search capability. However, in some cases, GWO converges to the local optimum and FWA converges slowly. In this paper, a new hybrid algorithm (named as FWGWO) is proposed, which fuses the advantages of these two algorithms to achieve global optima effectively. The proposed algorithm combines the exploration ability of the fireworks algorithm with the exploitation ability of the grey wolf optimizer (GWO) by setting a balance coefficient. In order to test the competence of the proposed hybrid FWGWO, 16 well-known benchmark functions having a wide range of dimensions and varied complexities are used in this paper. The results of the proposed FWGWO are compared to nine other algorithms, including the standard FWA, the native GWO, enhanced grey wolf optimizer (EGWO), and augmented grey wolf optimizer (AGWO). The experimental results show that the FWGWO effectively improves the global optimal search capability and convergence speed of the GWO and FWA.

Device-free localization (DFL) is an emerging wireless technique for estimating the location of target that does not have any attached electronic device. It has found extensive use in Smart City applications such as healthcare at home and hospitals, location-based services at smart spaces, city emergency response and infrastructure security. In DFL, wireless devices are used as sensors that can sense the target by transmitting and receiving wireless signals collaboratively. Many DFL systems are implemented based on received signal strength (RSS) measurements and the location of the target is estimated by detecting the changes of the RSS measurements of the wireless links. Due to the uncertainty of the wireless channel, certain links may be seriously polluted and result in erroneous detection. In this paper, we propose a novel nonlinear optimization approach with outlier link rejection (NOOLR) for RSS-based DFL. It consists of three key strategies, including: (1) affected link identification by differential RSS detection; (2) outlier link rejection via geometrical positional relationship among links; (3) target location estimation by formulating and solving a nonlinear optimization problem. Experimental results demonstrate that NOOLR is robust to the fluctuation of the wireless signals with superior localization accuracy compared with the existing Radio Tomographic Imaging (RTI) approach.

Recently, people have become more and more interested in wireless sensing applications, among which indoor localization is one of the most attractive. Generally, indoor localization can be classified as device-based and device-free localization (DFL). The former requires a target to carry certain devices or sensors to assist the localization process, whereas the latter has no such requirement, which merely requires the wireless network to be deployed around the environment to sense the target, rendering it much more challenging. Channel State Information (CSI)—a kind of information collected in the physical layer—is composed of multiple subcarriers, boasting highly fined granularity, which has gradually become a focus of indoor localization applications. In this paper, we propose an approach to performing DFL tasks by exploiting the uncertainty of CSI. We respectively utilize the CSI amplitudes and phases of multiple communication links to construct fingerprints, each of which is a set of multivariate Gaussian distributions that reflect the uncertainty information of CSI. Additionally, we propose a kind of combined fingerprints to simultaneously utilize the CSI amplitudes and phases, hoping to improve localization accuracy. Then, we adopt a Kullback−Leibler divergence (KL-divergence) based kernel function to calculate the probabilities that a testing fingerprint belongs to all the reference locations. Next, to localize the target, we utilize the computed probabilities as weights to average the reference locations. Experimental results show that the proposed approach, whatever type of fingerprints is used, outperforms the existing Pilot and Nuzzer systems in two typical indoor environments. We conduct extensive experiments to explore the effects of different parameters on localization performance, and the results demonstrate the efficiency of the proposed approach.

The classification of electrocardiograms (ECG) plays an important role in the clinical diagnosis of heart disease. This paper proposes an effective system development and implementation for ECG classification based on faster regions with a convolutional neural network (Faster R-CNN) algorithm. The original one-dimensional ECG signals contain the preprocessed patient ECG signals and some ECG recordings from the MIT-BIH database in this experiment. Each ECG beat of one-dimensional ECG signals was transformed into a two-dimensional image for experimental training sets and test sets. As a result, we classified the ECG beats into five categories with an average accuracy of 99.21%. In addition, we did a comparative experiment using the one versus rest support vector machine (OVR SVM) algorithm, and the classification accuracy of the proposed Faster R-CNN was shown to be 2.59% higher.

This paper exploits optimal position of the mobile sensor to improve the target tracking performance of wireless sensor networks and simultaneously localize both of the static sensor nodes and mobile sensor nodes when tracking the human motion. In our approach, mobile sensors collaborate with static sensors and move optimally to achieve the required detection performance. The accuracy of final tracking result is then improved as the measurements of mobile sensors have higher signal-to-noise ratios after the movement. Specifically, we can simultaneously localize the mobile sensor and static sensors position when localizing the human’s position based on augmented extended Kalman filters (EKF). In the algorithm, we develop a sensor movement optimization algorithm that achieves near-optimal system tracking performance. We also presented an sensor nodes management scheme in order to deduce the computation complexity when localizing the static sensor nodes. The effectiveness of our approach is validated by extensive simulations using the simulations.

Distributed estimation of Gaussian mixtures has many applications in wireless sensor network (WSN), and its energy-efficient solution is still challenging. This paper presents a novel diffusion-based EM algorithm for this problem. A diffusion strategy is introduced for acquiring the global statistics in EM algorithm in which each sensor node only needs to communicate its local statistics to its neighboring nodes at each iteration. This improves the existing consensus-based distributed EM algorithm which may need much more communication overhead for consensus, especially in large scale networks. The robustness and scalability of the proposed approach can be achieved by distributed processing in the networks. In addition, we show that the proposed approach can be considered as a stochastic approximation method to find the maximum likelihood estimation for Gaussian mixtures. Simulation results show the efficiency of this approach.

Collaborative target tracking is one of the most important applications of wireless sensor networks (WSNs), in which the network must rely on sensor scheduling to balance the tracking accuracy and energy consumption, due to the limited network resources for sensing, communication, and computation. With the recent development of energy acquisition technologies, the building of WSNs based on energy harvesting has become possible to overcome the limitation of battery energy in WSNs, where theoretically the lifetime of the network could be extended to infinite. However, energy-harvesting WSNs pose new technical challenges for collaborative target tracking on how to schedule sensors over the infinite horizon under the restriction on limited sensor energy harvesting capabilities. In this paper, we propose a novel adaptive dynamic programming (ADP)-based multi-sensor scheduling algorithm (ADP-MSS) for collaborative target tracking for energy-harvesting WSNs. ADP-MSS can schedule multiple sensors for each time step over an infinite horizon to achieve high tracking accuracy, based on the extended Kalman filter (EKF) for target state prediction and estimation. Theoretical analysis shows the optimality of ADP-MSS, and simulation results demonstrate its superior tracking accuracy compared with an ADP-based single-sensor scheduling scheme and a simulated-annealing based multi-sensor scheduling scheme.

Gas utilization ratio (GUR) is an important indicator that is used to evaluate the energy consumption of blast furnaces (BFs). Currently, the existing methods cannot predict the GUR accurately. In this paper, we present a novel data-driven model for predicting the GUR. The proposed approach utilized both the TS fuzzy neural network (TS-FNN) and the particle swarm algorithm (PSO) to predict the GUR. The particle swarm algorithm (PSO) is applied to optimize the parameters of the TS-FNN in order to decrease the error caused by the inaccurate initial parameter. This paper also applied the box graph (Box-plot) method to eliminate the abnormal value of the raw data during the data preprocessing. This method can deal with the data which does not obey the normal distribution which is caused by the complex industrial environments. The prediction results demonstrate that the optimization model based on PSO and the TS-FNN approach achieves higher prediction accuracy compared with the TS-FNN model and SVM model and the proposed approach can accurately predict the GUR of the blast furnace, providing an effective way for the on-line blast furnace distribution control.

This paper introduces a two-stage approach to the detection of people eating and/or drinking for the purposes of surveillance of daily life. With the sole use of wearable accelerometer sensor attached to somebody’s (man or a woman) wrists, this two-stage approach consists of feature extraction followed by classification. At the first stage, based on the limb’s three dimensional kinematics movement model and the Extended Kalman Filter (EKF), the realtime arm movement features described by Euler angles are extracted from the raw accelerometer measurement data. In the latter stage, the Hierarchical Temporal Memory (HTM) network is adopted to classify the extracted features of the eating/drinking activities based on the space and time varying property of the features, by making use of the powerful modelling capability of HTM network on dynamic signals which is varying with both space and time. The proposed approach is tested through the real eating and drinking activities using the three dimensional accelerometers. Experimental results show that the EKF and HTM based two-stage approach can perform the activity detection successfully with very high accuracy.

The concentration of alumina in the electrolyte is of great significance during the production of aluminum. The amount of the alumina concentration may lead to unbalanced material distribution and low production efficiency and affect the stability of the aluminum reduction cell and current efficiency. The existing methods cannot meet the needs for online measurement because industrial aluminum electrolysis has the characteristics of high temperature, strong magnetic field, coupled parameters, and high nonlinearity. Currently, there are no sensors or equipment that can detect the alumina concentration on line. Most companies acquire the alumina concentration from the electrolyte samples which are analyzed through an X-ray fluorescence spectrometer. To solve the problem, the paper proposes a soft sensing model based on a kernel extreme learning machine algorithm that takes the kernel function into the extreme learning machine. K-fold cross validation is used to estimate the generalization error. The proposed soft sensing algorithm can detect alumina concentration by the electrical signals such as voltages and currents of the anode rods. The predicted results show that the proposed approach can give more accurate estimations of alumina concentration with faster learning speed compared with the other methods such as the basic ELM, BP, and SVM.

Gas utilization ratio (GUR) is an important indicator used to measure the operating status and energy consumption of blast furnaces (BFs). In this paper, we present a soft-sensor approach, i.e., a novel online sequential extreme learning machine (OS-ELM) named DU-OS-ELM, to establish a data-driven model for GUR prediction. In DU-OS-ELM, firstly, the old collected data are discarded gradually and the newly acquired data are given more attention through a novel dynamic forgetting factor (DFF), depending on the estimation errors to enhance the dynamic tracking ability. Furthermore, we develop an updated selection strategy (USS) to judge whether the model needs to be updated with the newly coming data, so that the proposed approach is more in line with the actual production situation. Then, the convergence analysis of the proposed DU-OS-ELM is presented to ensure the estimation of output weight converge to the true value with the new data arriving. Meanwhile, the proposed DU-OS-ELM is applied to build a soft-sensor model to predict GUR. Experimental results demonstrate that the proposed DU-OS-ELM obtains better generalization performance and higher prediction accuracy compared with a number of existing related approaches using the real production data from a BF and the created GUR prediction model can provide an effective guidance for further optimization operation.

We advocate for and introduce LEARNSense, a framework for learning analytics using commodity wearable devices to capture learner’s physical actions and accordingly infer learner context (e.g., student activities and engagement status in class). Our work is motivated by the observations that: (a) the fine-grained individual-specific learner actions are crucial to understand learners and their context information; (b) sensor data available on the latest wearable devices (e.g., wrist-worn and eye wear devices) can effectively recognize learner actions and help to infer learner context information; (c) the commodity wearable devices that are widely available on the market can provide a hassle-free and non-intrusive solution. Following the above observations and under the proposed framework, we design and implement a sensor-based learner context collector running on the wearable devices. The latest data mining and sensor data processing techniques are employed to detect different types of learner actions and context information. Furthermore, we detail all of the above efforts by offering a novel and exemplary use case: it successfully provides the accurate detection of student actions and infers the student engagement states in class. The specifically designed learner context collector has been implemented on the commodity wrist-worn device. Based on the collected and inferred learner information, the novel intervention and incentivizing feedback are introduced into the system service. Finally, a comprehensive evaluation with the real-world experiments, surveys and interviews demonstrates the effectiveness and impact of the proposed framework and this use case. The F1 score for the student action classification tasks achieve 0.9, and the system can effectively differentiate the defined three learner states. Finally, the survey results show that the learners are satisfied with the use of our system (mean score of 3.7 with a standard deviation of 0.55).

Device-free localization (DFL) is becoming one of the new technologies in wireless localization field, due to its advantage that the target to be localized does not need to be attached to any electronic device. In the radio-frequency (RF) DFL system, radio transmitters (RTs) and radio receivers (RXs) are used to sense the target collaboratively, and the location of the target can be estimated by fusing the changes of the received signal strength (RSS) measurements associated with the wireless links. In this paper, we will propose an extreme learning machine (ELM) approach for DFL, to improve the efficiency and the accuracy of the localization algorithm. Different from the conventional machine learning approaches for wireless localization, in which the above differential RSS measurements are trivially used as the only input features, we introduce the parameterized geometrical representation for an affected link, which consists of its geometrical intercepts and differential RSS measurement. Parameterized geometrical feature extraction (PGFE) is performed for the affected links and the features are used as the inputs of ELM. The proposed PGFE-ELM for DFL is trained in the offline phase and performed for real-time localization in the online phase, where the estimated location of the target is obtained through the created ELM. PGFE-ELM has the advantages that the affected links used by ELM in the online phase can be different from those used for training in the offline phase, and can be more robust to deal with the uncertain combination of the detectable wireless links. Experimental results show that the proposed PGFE-ELM can improve the localization accuracy and learning speed significantly compared with a number of the existing machine learning and DFL approaches, including the weighted K-nearest neighbor (WKNN), support vector machine (SVM), back propagation neural network (BPNN), as well as the well-known radio tomographic imaging (RTI) DFL approach.

Energy-based source localization is an important problem in wireless sensor networks (WSNs), which has been studied actively in the literature. Numerous localization algorithms, e.g., maximum likelihood estimation (MLE) and nonlinear-least-squares (NLS) methods, have been reported. In the literature, there are relevant review papers for localization in WSNs, e.g., for distance-based localization. However, not much work related to energy-based source localization is covered in the existing review papers. Energy-based methods are proposed and specially designed for a WSN due to its limited sensor capabilities. This paper aims to give a comprehensive review of these different algorithms for energy-based single and multiple source localization problems, their merits and demerits and to point out possible future research directions.

The wearable inertial/magnetic sensor based human motion analysis plays an important role in many biomedical applications, such as physical therapy, gait analysis and rehabilitation. One of the main challenges for the lower body bio-motion analysis is how to reliably provide position estimations of human subject during walking. In this paper, we propose a particle filter based human position estimation method using a foot-mounted inertial and magnetic sensor module, which not only uses the traditional zero velocity update (ZUPT), but also applies map information to further correct the acceleration double integration drift and thus improve estimation accuracy. In the proposed method, a simple stance phase detector is designed to identify the stance phase of a gait cycle based on gyroscope measurements. For the non-stance phase during a gait cycle, an acceleration control variable derived from ZUPT information is introduced in the process model, while vector map information is taken as binary pseudo-measurements to further enhance position estimation accuracy and reduce uncertainty of walking trajectories. A particle filter is then designed to fuse ZUPT information and binary pseudo-measurements together. The proposed human position estimation method has been evaluated with closed-loop walking experiments in indoor and outdoor environments. Results of comparison study have illustrated the effectiveness of the proposed method for application scenarios with useful map information.

This paper proposes a human motion tracking approach for a daily life surveillance in a distributed wireless sensor network using ultrasonic range sensors. Because the human target often moves with high nonlinearity, the proposed approach applies the unscented Kalman filter (UKF) technique. A novel sensor node selection scheme at each time step considering both the tracking accuracy and the energy cost is presented. Experimental results in a real human motion tracking system show that the proposed approach can perform better tracking accuracy compared to the most recent human motion tracking scheme in the real testbed implementation.

The TDOA-based source localization problem in sensor networks is considered with sensor node location uncertainty. A total least squares (TLSs) algorithm is developed by a linear closed-form solution for this problem, and the uncertainty of the sensor location is formulated as a perturbation. The sensitivity of the TLS solution is also analyzed. Simulation results show its improved performance against the classic least squares approaches.