The structural damages of metallic structure components in many critical facilities and equipment may result in disasters that endanger human life. Existing Structural Health Monitoring (SHM) solutions commonly suffer from the limitations of bulky equipment, poor environmental adaptability, and high costs, which raise challenges for detection efficiency and large-scale multi-target monitoring. Radio Frequency Identification (RFID) sensing technology, featuring Non-Line-of-Sight (NLoS), flexible and pasteable, and easy deployment, show great promise for SHM. Recent studies have demonstrated the potential of RFID sensors for structural damage sensing including cracks, strain, and corrosion of metal structures, along with the analysis of parameters like crack width, structural deformation, and corrosion depth. This study provides a survey and in-depth analysis of recent technical progress in RFID sensor-based SHM. The main contributions include: (1) Classification of the novel sensing techniques and systems based on the functional model of RFID backscatter sensing; (2) Summarization of the common structural damage types and the feature extraction techniques of RFID sensing for SHM; (3) Survey of the recent progresses of the techniques, methods, and applications for RFID-based SHM; (4) Analysis of the challenges facing the state-of-the-art, including characterization and quantification of structure damage parameters and the impact of environmental factors, followed with an outlook of the future work. This study provides a timely reference for the innovation and practice of RFID sensing techniques in the field of SHM.

For the advantages of non-contact sensing, inventory identification, and cost-effective deployment, Radio Frequency Identification (RFID) localization has become a promising solution for some industrial Internet of Things (IoT) applications. However, its efficacy is often constrained by multi-antenna dependency, motion-induced phase distortions, and the inherent phase coupling between position and orientation, all of which hinder simultaneous detection of orientation and position in high accuracy. To overcome these issues, this investigation proposes a novel phase-decoupling model specifically designed for a single-antenna synthetic aperture radar (SAR) RFID system. The key contributions include: 1) Design and implementation of an adaptive dynamic phase compensation (ADPC) mechanism for decoupling motion-induced parameters from target backscatter signatures, effectively mitigating phase offsets caused by non-steady-state antenna trajectories; 2) Establishment of a novel phase-orientation model and a differential phase-adaptive peak detection (DPAPD) framework, which integrates differential measurements with threshold-optimized peak identification, achieving sub-degree angular resolution; 3) Development of a 3D SAR localization method incorporating phase decoupling and Particle Filter (PF) which achieves robust and consistent 3D localization with acceptable accuracy. This investigation provides a high-accuracy and cost-effective dynamic monitoring solution for RFID-based smart shelves, enabling advanced applications such as inventory tracking and tilt detection for fragile goods in automated warehouses.

Radio-Frequency Identification (RFID) positioning promises a prospective future for industrial automation and Industrial Internet of Things (IIoT) applications. However, the radio waves carry multiple parameters including position, orientation, and ambient environment factors, which raises challenges in simultaneous detection of position and orientation of product objects. This investigation proposes a simple RFID array-based position and orientation simultaneous detection technique for moving object in industrial chain. The main contributions of this investigation include: (1) Theoretical analysis and integrated model of position and orientation variation with the antenna parameters and interrogation variables in RF backscatter coupling-based sensing. (2) Development of an innovative simple RFID array-based phase separation technique with differential sensing, which determines the position-and orientation-induced phase without their mutual coupling impact. (3) Proposal of a simultaneous detection technique for moving objects’ position and orientation by integrating the Multiple Signal Classification (MUSIC) algorithm and hyperbolic positioning algorithm. In the experimental verification with a range from -75 cm to 75cm, the average error of position and orientation estimation is 4.29 cm and 4.89 degrees.

The development of the Internet of Things (IoTs) has empowered revolution in almost all walks of life. Although substantial effort has been made to bring IoT into manufacturing, there are still technical challenges to provide solutions of real-time pervasive multisensing, quality evaluation, and boundaryless information sharing, which put people at risk of deteriorated food products and food adulteration. In this article, we present an industrial IoT-based system for food product quality assessment and prediction. This research completes the real-time food quality assessment via radiofrequency identification (RFID) based multisensor fusion for the first time. Also, a novel concept of shelf-life prediction is proposed. The RFID-powered sensors provide a new idea for nondestructive food product and environment sensing. A five-layer architecture, considering sensing, controlling, communication, interfaces, and data analysis, is highlighted. A novel RFID metadata structure is first proposed to achieve multidimensional information traceability and global data sharing. Machine-learning-based multisensor fusion is proposed to provide the accurate quality assessment and prediction. The proposed system is implemented as a smart-shelf system to demonstrate its feasibility and advantages. The system is of great significance in improving food safety, reducing food waste, and providing powerful information support for food manufacturing line and supply chain management.

Without dependence on external anchors, the micro-electro-mechanical system inertial measurement unit (MIMU) allowing autonomous localization has promised great potential in wearable IoT applications, including kinematic analysis in sports, medical treatment, elderly care, and disaster rescue. However, the miniaturization of devices for unobtrusive sensing, algorithms minimizing the inherent accumulative errors of inertial devices, and the adaptivity of algorithms for various motion modalities are the key challenges. The removal of accumulative error in the continuous gait cycles with adaptive algorithms is a critical issue regarding localization accuracy, especially for low-cost devices. This investigation proposes an adaptive threshold-based zero-velocity update (ZUPT) algorithm to separate the timing of gait cycle phases and compensate for the residual velocity with a linear fitting approximation. The key contributions include: 1) a lightweight threshold-based zero-velocity detection algorithm to split the gait cycle phases of continuous walking; 2) a quaternion-based extended Kalman filter (EKF) algorithm to reduce the errors of the nonlinear operations for attitude prediction; 3) a linear fitting method for compensating the residual velocity in each gait cycle of continuous walking; and 4) the design of a miniature single-MIMU-based foot-mountable wearable device and the corresponding experimental studies to verify the proposed methods. Results show that the relative error is less than 3.0% for 2-D and 3-D trajectories, and the tests with different locomotion patterns demonstrate the adaptivity of the proposed algorithm compared to its peers. The results show that the presented techniques are capable of handling accumulative errors for low-cost MIMU-based systems with good adaptivity.

Depth map super-resolution (DMSR) is an effective solution to improve the quality of depth maps captured by low-cost depth sensors. Most existing methods introduce guidance from the RGB images of the same scene and achieve significant improvements. However, how to utilize the RGB information is still an open challenge because of the structure inconsistencies between RGB images and depth maps. In this letter, we present a novel convolutional neural network with deformable enhancement and adaptive fusion, termed DEAF-Net, to further improve the performance of DMSR. Specifically, we design a deformable convolution enhancement module, in which sufficient color features are used for enhancing depth features. An adaptively feature fusion module is exploited to improve the efficiency of fully connected feature fusion. Experimental results on two benchmark datasets demonstrate the effectiveness of the proposed method.

The Internet-of-Things (IoT) empowered technical revolution in manufacturing industry allows pervasive sensing and ubiquitous data access through the lifecycle of products. Although great efforts have been devoted to bridge the gap between industrial operations and information technologies, it still faces technical challenges to unobtrusively monitor the entire lifecycle of products via a boundary-less information flow for product lifecycle management in highly adaptive manufacturing. This investigation presents an IoT sensing and networking framework for seamless data integration and ubiquitous access in smart manufacturing, focusing on product identification, data modeling, interphase data integration, and ubiquitous data access. The highlights of this investigation are: 1) radio frequency identification (RFID) and virtual universal unique identifier dual identifier online item-specific data integration; 2) RFID-based online product object localization and unique identification; 3) object-centric manufacturing process modeling for interphase data integration; and 4) RFID/QR code encoding method for ubiquitous data sharing between product trading lifecycle phases. Finally, the implementation of presented methods in EU PickNPack food manufacturing production line is reported, and the practice has proved the feasibility and advantages.

Machine-to-machine (M2M) communication is a key enabling technology for the future industrial Internet of Things applications. It plays an important role in the connectivity and integration of computerized machines, such as sensors, actuators, controllers, and robots. The requirements in flexibility, efficiency, and cross-platform compatibility of the intermodule communication between the connected machines raise challenges for the M2M messaging mechanism toward ubiquitous data access and events notification. This investigation determines the challenges facing the M2M communication of industrial systems and presents a data-oriented M2M messaging mechanism based on ZeroMQ for the ubiquitous data access in rich sensing pervasive industrial applications. To prove the feasibility of the proposed solution, the EU funded PickNPack production line with a reference industrial network architecture is presented, and the communication between a microwave sensor device and the quality assessment and sensing module controller of the PickNPack line is illustrated as a case study. The evaluation is carried out through qualitative analysis and experimental studies, and the results demonstrate the feasibility of the proposed messaging mechanism. Due to the flexibility in dealing with hierarchical system architecture and cross-platform heterogeneity of industrial applications, this messaging mechanism deserves extensive investigations and further evaluations.

The integration of internet-of-things (IoT) technologies in the industry benefits digital manufacturing applications by allowing ubiquitous interaction and collaborative automation between machines. Online data collection and data interaction are critical for real-time decision making and machine collaborations. However, due to the specificity of digital manufacturing applications, the technical gap between IoT techniques and practical machine operation could hinder the efficient data interactions, collaborations between machines, and the effectiveness as well as the accuracy of itemized data collection. This investigation, therefore, identifies some major technical problems and challenges that current IoT-based digital manufacturing is facing, and proposes a method to bridge the technical gap for itemized product management. The highlights of this investigation are: 1) a data-oriented system architecture toward flexible data interaction between machines, 2) a customized machine-to-machine protocol for machine discovery, presence, and messaging, (3) flexible data structure and data presentation for interoperability, and (4) versatile information tracing approaches for product management. The proposed solutions have been implemented in PicknPack digital food manufacturing line, and achieved ubiquitous data interaction, online data collection, and versatile product information tracing methods have shown the feasibility and significance of the presented methods.

The continuous technical progress of the smartphone built-in modules and embedded sensing techniques has created chances for context-aware automation and decision support in home environments. Studies in this area mainly focus on feasibility demonstrations of the emerging techniques and system architecture design that are applicable to the different use cases. It lacks service customization strategies tailoring the computing service to proactively satisfy users’ expectations. This investigation aims to chart the challenges to take advantage of the dynamic varying context information, and provide solutions to customize the computing service to the contextual situations. This work presents a rule-based service customization strategy which employs a semantic distance-based rule matching method for context-aware service decision making and a Rough Set Theory-based rule generation method to supervise the service customization. The simulation study reveals the trend of the algorithms in time complexity with the number of rules and context items. A prototype smart home system is implemented based on smartphones and commercially available low-cost sensors and embedded electronics. Results demonstrate the feasibility of the proposed strategy in handling the heterogeneous context for decision making and dealing with history context to discover the underlying rules. It shows great potential in employing the proposed strategy for context-aware automation and decision support in smart home applications.