Human Activity Recognition (HAR), which aims to identify and classify human behaviors through multi-source sensing data, has become a long-standing research hotspot in ubiquitous computing, enabling a wide range of real-world applications. Focusing on Wearable HAR (WHAR), this survey provides a system-level synthesis organized around the life cycle of WHAR systems, collating fundamental theories, key technologies, and implementation details across data acquisition, model construction, and system deployment. Rather than treating sensing configuration, learning paradigm selection, evaluation protocol design, and deployment constraints as isolated topics, we emphasize their inherent interdependencies, demonstrating how early design choices propagate through the entire development pipeline to shape final system performance. We further establish a unified taxonomy of HAR paradigms based on data collection devices, clarifying the boundaries between vision-based, ambient sensor-based, and wearable sensor-based approaches to eliminate long-standing classification ambiguities. Additionally, we systematically review mainstream supervised WHAR models, emerging learning paradigms, deployment-oriented evaluation criteria, and cutting-edge directions including foundation models and Large Language Model (LLM)-based activity understanding. By integrating all aspects into a coherent life cycle framework, this survey aims to serve as a design-oriented reference for researchers and practitioners building accurate, efficient, and deployable WHAR systems.

Significant advancements in deep learning have been made possible by the utilization of large datasets, underscoring the critical importance of copyright protection. Adding meticulously designed perturbations to examples, making them unlearnable has become a crucial approach for safeguarding data copyright. Existing methods for creating unlearnable examples overlook the risk of data leakage, which can threaten data ownership. Thus, copyright protection in deep learning faces two main threats: illegal model training and malicious data leakage. We investigate that these two threats cannot be solved by straightforwardly combining existing availability attacks and watermarking techniques as their negative interaction effects. Therefore, in this paper, we propose a novel copyright protection mechanism for the aforementioned security concerns. Considering that the prevention of unauthorized model training requires powerful generalizability of unlearnable perturbations, we generate perturbations to induce the model to learn uncorrelated features of input images. It works by minimizing the mutual information of the input and output of the model. On the other hand, to eliminate the side impact of unlearnable perturbations on the watermark extraction, we design a dual extraction strategy by using two distinct watermark extractors. Extensive experiments on the image datasets ImageNet, CIFAR10, and Pets show that our proposed method could provide comprehensive copyright protection to images. The code is available at https://github.com/Yeah21/ReversibleUnlearnableExamples.

The proliferation of modern image editing tools has raised concerns about image manipulation, particularly regarding the potential to mislead the public and compromise privacy and security. Consequently, detecting and localizing tampered regions has become a critical research challenge. Traditional methods struggle with subtle manipulations, such as splicing, copy-move, and removal, which are often more discernible in the frequency domain than in the spatial domain. Additionally, the size imbalance between the tampered and background regions further complicates the detection process. To address these challenges, we propose DFFormer, an end-to-end network that leverages frequency feature differences and a dynamic token strategy for precise manipulation localization. DFFormer combines the Conventional Neural Network (CNN) and Transformer in a hybrid architecture with three key modules: the Adaptive Frequency Transformer (AFT), the Prototype Learning Module (PLM), and the Cascaded Progressive Token Fusion Head (CPTF-Head). AFT integrates high- and low-frequency components into self-attention via the Parallel Adaptive Frequency Attention (PAFA) block, enhancing tampering feature representation while preserving fine details. PLM employs KNN-based density peak clustering (DPC-KNN) and weighted token aggregation to optimize dynamic token reduction. The CPTF-Head adopts a hierarchical coarse-to-fine strategy to integrate multiscale features, thereby improving localization accuracy and edge refinement. Experiments demonstrate that DFFormer outperforms state-of-the-art models across four benchmark datasets and one real-world dataset, exhibiting superior generalization and robustness. The source code is publicly available at https://github.com/XiangGD/DFFormer.git.

The advent of extremely large antenna arrays and high-frequency signaling is expected to enable next-generation integrated sensing and communication (ISAC) networks to predominantly operate in the near-field region. Due to the dual influence of distance and angle on wave propagation characteristics in the near-field region, accurately modeling these characteristics remains a critical challenge. Motivated by the potential of large language models (LLMs) in angle prediction and distance estimation, an LLM-enhanced multi-objective optimization problem (MOOP) is developed to accurately capture the dependence of the channel on both the angular position and distance. The formulated LLM-enhanced MOOP framework is decomposed into a series of sub-problems, which can balance spectral efficiency for communication and localization accuracy for sensing. To overcome the computational and energy challenges associated with LLMs, a gray wolf optimization (GWO)-based algorithm is integrated as black-box search operator with LLM-specific prompt engineering to solve these sub-problems. Numerical results demonstrate that the proposed LLM-GWO scheme achieves an trade-off between communication and sensing performance, outperforming baseline approaches in terms of both Pareto front quality and convergence.

Traditional traffic anomaly-based attack detection methods in Zero-trust Networks (ZTN) suffer from inherent security vulnerabilities, as they neglect considerations regarding their security defenses. Compromising the attack detection model itself can result in the breakdown of normal attack detection capabilities. Ensuring the security of the attack detection model during runtime presents a novel challenge. To address these shortcomings, we propose a novel attack detection model, termed Security within Security: Attack Detection Model with Defenses Against Attacks Capability for Zero-Trust Networks (SWS), aimed at enhancing the security of ZTN. SWS focuses on achieving attack detection in non-secure detection environments, to maintain its detection capability even when under attack. By employing a soft thresholding method, SWS adapts to the dynamic changes in network traffic, thus reducing the interference of attack signals. The incorporation of an attention mechanism enables SWS to concentrate on analyzing the most indicative traffic features of attack behavior. Additionally, we integrate Residual Networks (ResNet) and Bidirectional Long Short-Term Memory (BiLSTM) to enhance the robustness of identifying complex network attack behaviors. The effectiveness of the SWS is validated through ablation studies, model comparisons, experiments conducted over different training epochs, and experiments conducted on various components of the dataset. Experimental results demonstrate that compared to existing attack detection models, SWS achieves improvements in detection accuracy and recall rate by 13.4% and 10.6%, respectively, while reducing the False Positive Rate (FPR) by 16.9%.

Trajectory prediction is a crucial challenge in autonomous vehicle motion planning and decision-making techniques. However, existing methods face limitations in accurately capturing vehicle dynamics and interactions. To address this issue, this article proposes a novel approach to extracting vehicle velocity and acceleration, enabling the learning of vehicle dynamics and encoding them as auxiliary information. The VDI-LSTM model is designed, incorporating graph convolution and attention mechanisms to capture vehicle interactions using trajectory data and dynamic information. Specifically, a dynamics encoder is designed to capture the dynamic information, a dynamic graph is employed to represent vehicle interactions, and an attention mechanism is introduced to enhance the performance of LSTM and graph convolution. To demonstrate the effectiveness of our model, extensive experiments are conducted, including comparisons with several baselines and ablation studies on real-world highway datasets. Experimental results show that VDI-LSTM outperforms other baselines compared, which obtains a 3% improvement on the average RMSE indicator over the five prediction steps.

A bike-sharing system (BSS) is easily unbalanced due to the uncertainty of user demand at each bike station during the day, which appeals for an effective bike reposition solution based on the accurate prediction of user demand. However, there is a discrepancy between the bike pickup/drop-off record (satisfied demand) and the user’s first choice of origin/destination stations (i.e., true user demand) since the BSS cannot capture the unsatisfied user demand (i.e., abandoned rentals and transferred rentals/returns) that occurs at either empty stations (failed rentals) or full stations (failed returns). To efficiently rebalance the BSS, this paper focuses on accurately forecasting the true user demand of the BSS. First, we extract the spatial-temporal features of bike usage and establish a spatio-temporal model for true user demand prediction. Then, an XGBoost-based three-stage prediction approach is proposed to accurately predict the true user demand including the station clustering, the system record rectification, and the true user demand prediction. The real data from the Citi Bike in New York is applied to verify the proposed method and the experimental results demonstrate that the proposed approach outperforms the existing methods.

The security of the Internet of Vehicles (IoV) has always been a concern. The constantly changing IoV data under varying traffic conditions made it unsuitable for the IoV to adopt traditional anti-attack techniques. In the absence of protections, attackers can use in-car communication as a target to compromise the safety of passengers, hence the instancy to detect the security state of the IoV. However, currently available solutions require modifications to the original hardware of the IoV and are therefore very limited in applicability. In this paper, we propose a security assessment method for IoV based on Microcontroller Unit (MCU) chip temperature, called SAMCT. Specifically, we first record the MCU chip temperatures of IoV device in different security states and analyze the relationship between them. Second, the fingerprint dataset is built using the temperature residuals. Third, to forecast the security standing of IoV devices, an integration regression model based on Self-Encoders is suggested. Lastly, in order to facilitate the effectiveness of the SAMCT, a Cloud-Edge-End framework is designed with the technology of model adaptive partitioning. Results from the experiments, which were carried out on the Raspberry Pi 4B and Stm32 hardware platforms, demonstrate that the Mean Squared Error (MSE) of the SAMCT is only 0.00104 and that the execution efficiency improvement under the Cloud-Edge-End framework is significant.Note to Practitioners—This paper was inspired by security concerns in Internet of Vehicles communication systems. The core of this work is to provide a novel security assessment method for IoV devices based on MCU temperature, which can detect the security status of IoV devices in real-time without modifying the original hardware. To this end, the different skills from scheme design to detection and validation are explained. One crucial part of this work is to regard the MCU temperature of the IoV device as a security reference and fully integrate the critical techniques in deep learning. In addition, the proposed scheme is universal and can be applied to various scenarios such as the autonomous driving and the industrial internet of things.

The continual progression and expanding applications of Hyperspectral Imaging (HSI) technology necessitate the development of lightweight HSI classification models that are capable of lifelong learning. However, the computationally-demanding task of training and updating HSI classification models, exacerbated by the substantial number of trainable parameters in feature extractors, remains a substantial challenge. This paper proposes an Attention-based Reservoir Computing (ARC) model to overcome these hurdles. The ARC model utilizes a cross-slicing operation to generate multi-directional inputs, treating the HSI dataset as spatial sequence for processing within a reservoir. For every target pixel, four spatial sequences from various directions are introduced into the reservoir, generating four corresponding outputs. A voting mechanism then evaluates these outputs to yield the final prediction. Additionally, we design an attention-based leaky function for reservoir computing to capture the spatial correlation inherent in HSI data accurately. The attention-based leaky function enables the reservoir state to weigh less on the pixels outside the region of interest (ROI) and have a longer memory for pixels inside the ROI. The ARC was tested on widely used HSI datasets: Indian Pines, PaviaU, and Salinas. It demonstrated competitive lightweight classification performance against state-of-the-art lightweight models by maintaining comparable training time while achieving superior accuracy. Furthermore, the model’s lifelong learning accuracy also showed superior performance compared to existing lifelong learning models, with a one thousand times reduction of parameter-to-be-updated. This work makes the ARC model an effective contender for HSI classification tasks, excelling in both lightweight classification and lifelong learning capacities. The source codes are available publicly at: https://github.com/Waterman-Ann/ARC.

Copy–move forgery is one of the most frequently used methods for producing fake digital images. Current algorithms for copy–move forgery detection (CMFD) cannot combine high accuracy and fast speed. Motivated by the observation, we propose a novel CMFD algorithm whose workflow is as follows. First, we use a keypoint-extraction method with the lowest contrast threshold to extract more keypoints from the input image. Second, a new technique, gradient-hash matching, finds pairs of similar keypoints quickly and effectively using a hash table, where the hash value is computed using gradients of keypoints. Subsequently, a new method called simplified cluster-based filtering exploits the density pattern of keypoints in the copy–move regions to remove false matching keypoint pairs. Finally, image matting is applied to indicate the forgery regions vividly. Extensive experiments show that not only the new algorithm is better than the state-of-the-art algorithms in terms of computation correctness, but also its computation time is drastically less. Commonly only about half time is needed. The relative time saving is even higher when images are larger. Different algorithms modules are compared through experiments to choose the best combination.

Disassembly is the first phase to demanufacture end-of-life (EOL) products that are separated into parts/components for recovery. The quality conditions of EOL products are highly uncertain, which would result in some uncertain information during the disassembly process, e.g., the disassembly time and recovering revenue of each subassembly. It is quite challenging to determine the optimal/near-optimal disassembly solutions under uncertain information. This article studies uncertain disassembly planning (UDP) and proposes a self-adaptive learning approach to quickly identify the near-optimal disassembly solutions. First, we model the UDP by extending Petri nets, where not only disassembly operations but also EOL options of each subassembly are represented in the extended Petri Net. Based on the UDP model, we develop the self-adaptive learning approach, which integrates an approximation procedure for estimating uncertain disassembly information, aQ-learning algorithm for training disassembly samples, and a heuristic method for selecting the best disassembly solution. Finally, a hybrid Li-ion battery pack of Audi A3 Sportback e-tron is selected as the case study and applied to test the proposed self-adaptive learning approach. The experimental results demonstrate that our proposed method can efficiently find a better disassembly solution than the existing disassembly solution within 200 trainings in the case study.

Physics-informed learning for industrial Internet is essential especially to safety issues. Consequently, various methods have been developed to conduct Industrial Internet of Things (IIoT) intrusion detection. However, the conventional methods usually require the help of auxiliary equipment (e.g., spectrum analyzers, log-periodic antennas), which proves to be unsuitable for general IIoT systems due to their poor versatility. Facing the dilemma mentioned above, this article proposes a microcontroller unit (MCU) chip temperature fingerprint informed machine learning method, called MTID, for IIoT intrusion detection. Specifically, first, the node's MCU temperature sequence is recorded and the relationship between the temperature sequence and the computational complexity of the node is analyzed. Then, we calculate the temperature residuals and construct a temperature residuals dataset. Finally, to identify the security status of the nodes, a self-encoder-based intrusion detection model is constructed. Furthermore, to ensure the model's applicability under the diversified deployment environment of IIoT systems, an online incremental training method is developed and applied. In the end, we use the Raspberry Pi 4B for experimental analysis when testing the performance of MTID. The results show that the accuracy of MTID for intrusion detection reaches 89%, which also demonstrates the feasibility of the intrusion detection method based on MCU temperature.

Due to the limited computing power of Industrial Internet of Things (IIoT), it is impossible to port traditional attack resistance methods to run in IIoT systems. Currently, various methods have been developed to conduct security assessment for IIoT systems. However, these methods require modification of the original hardware of the IIoT system, so they are not universally applicable. In this paper, we propose a Non-intrusive Security Estimation Method (NSEM) for IIoT systems based on common attribute of IIoT devices. In the NSEM, we firstly record the common attribute (i.e. MCU chip temperatures) in different security states, and construct the temperature fingerprint dataset. Then, a Self-Encoder-based integration regression model is proposed to predict the security status of IIoT devices. Finally, we design a Cloud-Edge-End framework with model adaptive partitioning technology to support the efficient execution of the NSEM method. The experiments are conducted on Raspberry Pi 4B platforms. The results show that the Mean Squared Error (MSE) of the NSEM is only 0.001, and the Cloud-Edge-End framework can effectively improve execution efficiency.

Selecting reliable beacon nodes plays a significant role in relocalizing unknown nodes in a wireless sensor network. When the position of a beacon node is drifted or is spoofed, it becomes an unreliable beacon node, which would lead to a large relocalization deviation of unknown nodes in its neighbor. However, when selecting reliable beacon nodes, most relocalization algorithms only screen either drifting beacon nodes or malicious beacon nodes whose position is drifted or spoofed. This article proposes an algorithm that can simultaneously screen drifting beacon nodes and malicious beacon nodes. The algorithm is divided into four steps. First, three indicators are introduced, where two are for describing position drifting and one is for describing position spoofing. Second, the entropy method is used to weight the contributions of three indicators. Third, a technique for order preference by similarity to an ideal solution is used to construct a reliability evaluation model. Finally, using the reliability evaluation model select reliable beacon nodes. Experimental results illustrate that the detection accuracy of drifting beacon nodes and malicious beacon nodes of the proposed algorithm is 7.5% and 8.2% higher than that of the state-of-the-art algorithms, respectively.

Task scheduling on edge computing servers is an important issue that affects user experience. Existing scheduling methods require centralized control to achieve the best overall performance. However, it is impractical to force all users to act according to centralized control. We propose a distributed edge computing server task scheduling model based on game theory. Our method comprehensively considers the link quality from the mobile device to the server and the server's computing resource allocation when selecting edge computing servers, and achieves a balance between link quality and computing resources. Once the Nash equilibrium is reached, our model can provide different QoS for users of different priorities. Acceleration methods are proposed to achieve the Nash equilibrium faster. The simulation results show that the proposed model can provide differentiated services while optimizing the scheduling of computing resources, and ensure that the algorithm achieves an approximate Nash equilibrium in polynomial time.

Frequently changing pseudonyms is one of commonly accepted approaches to protect location privacy in Vehicular Ad hoc NETworks (VANETs), but most pseudonyms change schemes based on vehicular individual behavior are inefficient and the level of anonymity provided by these schemes is difficult to be analyzed. In this paper, we discuss how Cooperative Pseudonym Change (CPC) scheme enhances anonymity and develop an approximate analysis model for the scheme in one-dimensional VANETs to quantify the level of anonymity where vehicles are uniformly distributed on road. The accuracy of this model is verified by a set of simulations, and the results show that the enhanced anonymity rate of CPC scheme over Non-Cooperative Pseudonym Change (NCPC) scheme has a trend of first increase and then decrease with the increase of the average number of neighbors of the target vehicle.