In recent years, WiFi-based device-free localization (DFL) has attracted attentions due to the rapid development of location-based applications. The localization performance of data-driven DFL models highly relies on the quality of the fingerprint. However, it is difficult to obtain high-quality fingerprints in cluttered environments due to various uncertainties, such as environmental dynamics. To mitigate effects of uncertainties, this article proposes a variance-constrained local–global modeling method to enhance the localization performance. To be specific, the collected channel state information data from a specific environment are first divided into several groups depending on their statistical characteristics using the clustering method, and the local–global modeling mechanism is then implemented based on the extreme learning machine, a kind of noniterative single hidden layer feedforward neural network, to achieve the good representation of the whole environment. During the local–global modeling process, the proposed method is to simultaneously minimize the output weights of the neural network, training errors of the DFL model, and intragroup variances of the grouped data, which makes the created DFL model robust in cluttered environments and not sensitive to uncertainties. Comprehensive experiments in several scenarios, including different indoor environments, device positions and heights, target's orientations, and body shapes, are performed. Experimental results indicate that the proposed method could achieve better localization performance than selected baseline methods, demonstrating the effectiveness of the proposed variance-constrained local–global modeling mechanism in DFL.

Ubiquitous WiFi signals not only provide fundamental communications for a large number of Internet of Things devices, but also enable to estimate target’s location in a contactless manner. However, most of the existing device-free localization (DFL) methods only utilize the time dynamics of the received WiFi signals, leading to inaccurate DFL in the cluttered indoor environments. Because different layouts of environments and deployments of WiFi devices cause the different mathematical distributions of the data collected from the cluttered indoor environments. In this paper, a multiple kernel representation based extreme learning machine (ELM) is proposed, named integrated multiple kernel ELM (IMK-ELM), for strengthening the localization performance in the cluttered indoor environments utilizing spatiotemporal information. In the proposed IMK-ELM-based DFL, the whole dataset is first divided into several subsets depending on their mathematical distributions through K-means clustering algorithm, and then corresponding number of local DFL models are built for all the subsets to capture both the time dynamics and spatial properties of the data. Finally, a global DFL model is achieved by seamlessly integrating all the local DFL models due to the consistency mechanism. In addition, Fresnel zone sensing theory is utilized for helping understand and explain the essence of indoor DFL. Comprehensive experiments indicate that the proposed IMK-ELM-based DFL outperforms state-of-the-arts in the cluttered indoor environments.

Vital signs, such as respiration rate (RR) and heart rate (HR), are essential for human health status assessment. The radar could detect RR and HR in a non-contact manner by sensing the repetitive chest wall movement caused by cardiopulmonary activity, which is attractive due to the more comfortable experience and better privacy protection. However, as the chest wall displacement caused by heartbeat is much smaller than that caused by respiration, the weak heartbeat compo-nent may be easily overwhelmed by respiration harm-onics, noise and clutter, making it difficult to achieve accurate and reliable HR detection. In order to tackle this challenge, in this paper, we design mmRH, a system can accurately estimate RR as well as HR using a frequency modulated continuous wave (FMCW) millimeter-wave (mm-wave) radar, which consists of four functional modules. In the vital sign signal extraction module, the range bin corresponding to the body part with strong vital signs is determined to extract the vital sign signal, eliminating the clutter interference. The differential enhancement module is proposed to enhance the heartbeat component by the first order temporal difference, reducing the impacts of respiration harmonics and noise on HR estimation. In the signal decomposition module, the respiration and heartbeat signals are separated via wavelet packet decomposition (WPD) to further suppress respiration harmonics and high-frequency noise of the heartbeat signal. In the vital sign rate reconstruction module, the sparse spectrum reconstruction (SSR) of vital sign rates is mapped to an adaptive filter and zero attracting sign exponentially forgetting least mean square (ZA-SEFLMS) algorithm is proposed to achieve high-resolution sparse spectrum for accurate and reliable RR and HR estimation. The effects of different settings, including the distance between the human subject and the radar, the sensor position and the user heterogeneity, are investigated by extensive experiments. The results indicate that the designed mmRH could effectively suppress the respiration harmonics, noise and clutter interference, and bring a significant improvement in RR and HR detection accuracy compared with existing methods.

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.

In this paper, an intelligent data-driven optimization scheme is proposed for finding the proper burden surface distribution, which exerts large influences on keeping blast furnace running smoothly in energy-efficient state. In the proposed scheme, production indicators prediction models are firstly developed using kernel extreme learning machine algorithm. To heel, burden surface decision is presented as a multi-objective optimization problem for the first time and solved by a modified two-stage intelligent optimization strategy to generate the initial setting values of burden surface. Furthermore, considering the existence of approximation error of the created prediction models, feedback compensation is implemented to enhance the reliability of the results, in which, an improved association rules mining method is developed to find the corrected values to compensate the initial setting values. Finally, we apply the proposed optimization scheme to determine the setting values of burden surface using actual data, and experimental results illustrate its effectiveness and feasibility.

Human activity recognition (HAR) is a key technology in the field of human–computer interaction. Unlike systems using sensors or special devices, the WiFi channel state information (CSI)-based HAR systems are noncontact and low cost, but they are limited by high computational complexity and poor cross-domain generalization performance. In order to address the above problems, a reconstructed WiFi CSI tensor and deep learning based lightweight HAR system (Wisor-DL) is proposed, which firstly reconstructs WiFi CSI signals with a sparse signal representation algorithm, and a CSI tensor construction and decomposition algorithm. Then, gated temporal convolutional network with residual connections is designed to enhance and fuse the features of the reconstructed WiFi CSI signals. Finally, dendrite network makes the final decision of activity instead of the traditional dense layer. Experimental results show that Wisor-DL is a lightweight HAR system with high recognition accuracy and satisfactory cross-domain generalization ability.