Atrial fibrillation (AF) is one of the most common sustained chronic cardiac arrhythmia in elderly population, associated with a high mortality and morbidity in stroke, heart failure, coronary artery disease, systemic thromboembolism, etc. The early detection of AF is necessary for averting the possibility of disability or mortality. However, AF detection remains problematic due to its episodic pattern. In this paper, a multiscaled fusion of deep convolutional neural network (MS-CNN) is proposed to screen out AF recordings from single lead short electrocardiogram (ECG) recordings. The MS-CNN employs the architecture of two-stream convolutional networks with different filter sizes to capture features of different scales. The experimental results show that the proposed MS-CNN achieves 96.99% of classification accuracy on ECG recordings cropped/padded to 5 s. Especially, the best classification accuracy, 98.13%, is obtained on ECG recordings of 20 s. Compared with artificial neural network, shallow single-stream CNN, and VisualGeometry group network, the MS-CNN can achieve the better classification performance. Meanwhile, visualization of the learned features from the MS-CNN demonstrates its superiority in extracting linear separable ECG features without hand-craft feature engineering. The excellent AF screening performance of the MS-CNN can satisfy the most elders for daily monitoring with wearable devices.

With the rapid development of wearable IoT devices, a significant amount of sensitive/private information collected by them poses a considerable challenge to the security of the wearable IoT devices. The accelerometer-based gait recognition is considered as an emerging and fast-evolving technology in security and access control fields and has achieved outstanding performance at certain fixed walking speeds. However, the gait recognition performance of the above technology deteriorates dramatically when the walking speed varies. To address this issue, both the speed-adaptive gait cycle segmentation method and individualized matching threshold generation method were proposed in this paper. Furthermore, the contrast experiments were conducted on the ZJU-GaitAcc public dataset sampled from five different body locations and the self-collected dataset sampled at various walking speeds. The experimental results indicated the average gait recognition and user authentication rates of 96.9% and 91.75%, respectively. As compared to the available state-of-the-art methods based on the fixed walking speeds and constant thresholds, the proposed method improved the gait recognition by 25.8% and user authentication by 21.5%.

Wearable sensors-based gait recognition is an effective method to recognize people’s identity by recognizing the unique way they walk. Recently, the adoption of deep learning networks for gait recognition has achieved significant performance improvement and become a new promising trend. However, most of the existing studies mainly focused on improving the gait recognition accuracy while ignored model complexity, which make them unsuitable for wearable devices. In this study, we proposed a lightweight attention-based Convolutional Neural Networks (CNN) model for wearable gait recognition. Specifically, a four-layer lightweight CNN was first employed to extract gait features. Then, a novel attention module based on contextual encoding information and depthwise separable convolution was designed and integrated into the lightweight CNN to enhance the extracted gait features and simplify the complexity of the model. Finally, the Softmax classifier was used for classification to realize gait recognition. We conducted comprehensive experiments to evaluate the performance of the proposed model on whuGait and OU-ISIR datasets. The effect of the proposed attention mechanisms, different data segmentation methods, and different attention mechanisms on gait recognition performance were studied and analyzed. The comparison results with the existing similar researches in terms of recognition accuracy and number of model parameters shown that our proposed model not only achieved a higher recognition performance but also reduced the model complexity by 86.5% on average.

Accurate assessment of physical fatigue is crucial to preventing physical injury caused by excessive exercise, overtraining during daily exercise and professional sports training. However, as a subjective feeling of an individual, physical fatigue is difficult for others to objectively evaluate. Heart rate variability (HRV), which is derived from electrocardiograms (ECG) and controlled by the autonomic nervous system, has been demonstrated to be a promising indicator for physical fatigue estimation. In this paper, we propose a novel method for the automatic and objective classification of physical fatigue based on HRV. First, a total of 24 HRV features were calculated. Then, a feature selection method was proposed to remove useless features that have a low correlation with physical fatigue and redundant features that have a high correlation with the selected features. After feature selection, the best 11 features were selected and were finally used for physical fatigue classifying. Four machine learning algorithms were trained to classify fatigue using the selected features. The experimental results indicate that the model trained using the selected 11 features could classify physical fatigue with high accuracy. More importantly, these selected features could provide important information regarding the identification of physical fatigue.

Indoor location and navigation technologies are crucial for healthcare, security and other location-based services. Wi-Fi and inertial sensors have become mainstream indoor localization technologies for wearable device platforms due to simple deployment and low cost. This study proposes an extended Kalman filtering (EKF)-based multimodal sensor fusion algorithm for indoor localization, combining Wi-Fi fingerprint and inertial measurement unit (IMU) data to provide accurate and continuous pedestrian localization. The main contributions of this work are threefold. First, a Wi-Fi fingerprint data augmentation method based on access point (AP) location sorting is proposed and a regression network model with a convolutional denoising autoencoder for WiFi-based indoor localization (CDAELoc) is designed to improve the robustness. Second, a dual-branch deep inertial odometry (DbDIO) network model for IMU-based indoor localization is introduced, consisting of two branches with various convolutional kernel sizes to extract features at different scales. Finally, an EKF-based Wi-Fi and inertial odometry (WIO-EKF) fusion localization system is presented, utilizing the predicted results from the proposed CDAELoc and DbDIO models as the system observations and mitigating the initial heading error of DbDIO. The proposed models are applied to the UJIIndoorLoc, RoNIN public data sets and self-collected data set. Experimental results prove that the proposed CDAELoc model outperforms other Wi-Fi localization models, reducing the average positioning error (APE) by 12.5%. The proposed DbDIO model achieves higher accuracy and requires fewer model parameters than any other deep inertial odometry model. Finally, the APE of WIO-EKF is lower than those of CDAELoc and DbDIO by 34% and 42%.

With the fast development of wearable IoT devices, their applications are becoming more and more pervasive, ranging from social networking, payment, and navigation to health and activity monitoring. The security of the communication between these devices is essential to protect the transmitted sensitive information from tampering and eavesdropping. With the integration of accelerometers into wearable IoT devices, the gait-based biometric cryptography technology has emerged as a data securing tool for wearables. This paper proposes a light-weight noise-based group key generation method, which utilizes the noise signals imposed on the raw acceleration signals to generate an M-bit key with high randomness and bit generation rate. Moreover, a signed sliding window coding (SSWC)-based common feature extraction method was designed to extract the common feature for sharing the generated M-bit key among devices worn on different body parts. Finally, a fuzzy vault-based group key distribution system was implemented and evaluated using a public dataset. The performed comprehensive analysis of the proposed key generation and distribution method proved that the binary keys generated via the introduced noise-based procedure have high entropy and can pass both the NIST and Dieharder statistical tests with high efficiency. The experimental results obtained prove the robustness of the proposed SSWC-based common feature extraction method in terms of the similarity and discriminability of intra-and inter-class features, respectively.