The rapid advancement of pre-trained language models (PLMs) has demonstrated promising results for various code-related tasks. However, their effectiveness in detecting real-world vulnerabilities remains a critical challenge. While existing empirical studies evaluate PLMs for vulnerability detection (VD), they suffer from data leakage, limited scope, and superficial analysis, hindering the accuracy and comprehensiveness of evaluations. This paper begins by revisiting the common issues in existing research on PLMs for VD through the evaluation pipeline. It then proceeds with an accurate and extensive evaluation of 18 PLMs, spanning model parameters from millions to billions, on high-quality datasets that feature accurate labeling, diverse vulnerability types, and various projects. Specifically, we compare the performance of PLMs under both fine-tuning and prompt engineering, assess their effectiveness and generalizability across various training and testing settings, and analyze their robustness to perturbations such as code normalization, abstraction, and semantic-preserving transformations.

This paper proposes FDALLM-Small, a unified and lightweight RF sensing framework that integrates Functional Data Analysis (FDA) with parameter efficiently tuned small language models. By transforming raw RF measurements into smooth and structured functional embeddings and encoding them into standardized functional prompts, the framework enables compact LLMs to perform classification and localization tasks with strong accuracy and robustness. Through LoRA based fine tuning, small LLMs effectively learn discriminative RF patterns while updating only a tiny fraction of model parameters, making the approach highly efficient and suitable for on device deployment. Experiments on the XRF55 and AdaRF datasets demonstrate that the FDA–prompting pipeline substantially boosts model performance, allowing small LLMs to surpass conventional deep learning baselines and approach the accuracy of large API based LLMs without relying on cloud computation. A scaling study further shows that smaller models consistently offer the best performance–efficiency trade offs, highlighting the intrinsic compatibility between FDA representations and compact architectures. These results confirm the practicality of FDALLM-Small as an edge friendly and computationally efficient solution for real world RF sensing applications.

Human activity recognition (HAR) requires extracting accurate spatial-temporal features with human movements. A mmWave radar point cloud-based HAR system suffers from sparsity and variable-size problems due to the physical features of the mmWave signal. Existing works usually borrow the preprocessing algorithms for the vision-based systems with dense point clouds, which may not be optimal for mmWave radar systems. In this work, we proposed a graph representation with a discrete dynamic graph neural network (DDGNN) to explore the spatial-temporal representation of human movement-related features. Specifically, we designed a star graph to describe the high-dimensional relative relationship between a manually added static center point and the dynamic mmWave radar points in the same and consecutive frames. We then adopted DDGNN to learn the features residing in the star graph with variable sizes. Experimental results demonstrated that our approach outperformed other baseline methods using real-world HAR datasets. Our system achieved an overall classification accuracy of 94.27%, which gets the near-optimal performance with a vision-based skeleton data accuracy of 97.25%. We also conducted an inference test on Raspberry Pi 4 to demonstrate its effectiveness on resource-constraint platforms. We provided a comprehensive ablation study for variable DDGNN structures to validate our model design. Our system also outperformed three recent radar-specific methods without requiring resampling or frame aggregators.

While supervised deep neural networks (DNNs) have proven effective for device authentication via radio frequency (RF) fingerprinting, they are hindered by domain shift issues and the scarcity of labeled data. The success of large language models has led to increased interest in self-supervised pre-trained models (PTMs), which offer better generalization and do not require labeled datasets, potentially addressing the issues mentioned above. However, the inherent vulnerabilities of PTMs in RF fingerprinting remain insufficiently explored. In this paper, we unveil the potential threat by thoroughly investigating data-free backdoor attacks on such PTMs for RF fingerprinting, focusing on a practical scenario where attackers lack access to downstream data, label information, and training processes. To realize the backdoor attack, we carefully design a set of triggers and predefined output representations (PORs) for the PTMs. By mapping triggers and PORs through backdoor training, we can implant backdoor behaviors into the PTMs, thereby introducing vulnerabilities across different downstream RF fingerprinting tasks without requiring prior knowledge. Extensive experiments demonstrate the wide applicability of our proposed backdoor attack to various input domains, protocols, and PTMs. Furthermore, we explore potential detection and defense methods, illustrating the difficulty of fully safeguarding against our proposed data-free backdoor attack.

The cybersecurity market is rapidly growing in the face of ubiquitous vulnerabilities, with vendors delivering solutions ranging from basic software-based approaches to high-end services for diverse business needs. The sustainability of a cybersecurity business relies on its ability to update its solutions to address ever-changing cyber threats continuously. It requires a substantial customer base for data collection and service enhancement, which, in turn, can attract more customers. Therefore, a novel business model that acknowledges this positive network effect and provides incentives for potential customers is crucial to ensure the success of the cybersecurity business. In this paper, we develop an analytical framework with optimization and a Stackelberg game approach to model the cybersecurity market scenario for a cybersecurity vendor offering both a basic software-based service and a premium service. We delve into the market evolution and characterize the conditions under which the dynamic market converges to a unique equilibrium. The optimal pricing strategy for the vendor is analyzed to leverage the network effects for profit maximization.

In communication network management, mobile traffic prediction is vital for ensuring efficient system operation. Despite considerable progresses in applying neural networks for traffic prediction, traditional models often struggle to handle high-dimensional and time-dependent data. This paper addresses these challenges by proposing a novel framework that constructs prompts to enhance the predictive ability of large language models (LLMs) and their understanding of traffic data. Specifically, we leverage functional data analysis (FDA), a superior technique to traditional methods, to preprocess traffic data and extract features. Through extensive experiments on various LLMs with a real-world dataset, we validate the effectiveness and scalability of our proposed method, with performance improvements of up to 23.53 % and 21.34 % in mean squared error (MSE) and mean absolute error (MAE), respectively. Our results indicate a significant advance in predictive performance, providing a promising approach for future traffic data analysis.

Radio frequency (RF) fingerprinting is a technique used to identify a wireless device based on its specific and unique hardware characteristics. In recent years, deep learning has been utilized for RF fingerprinting due to its superiority in feature extraction and higher classification accuracy. However, one major challenge of deep learning-based RF fingerprinting is that wireless signals are highly sensitive to environmental conditions, causing the device fingerprints captured in one environment to not transfer well to another. Hence, deep learning models are found to perform well in the same condition but lose their ability to classify devices in the new condition. In this paper, we examine three transfer learning techniques to mitigate the domain shift problem in RF fingerprinting and compare them with two well-defined baselines. The three RF fingerprinting datasets under various scenarios are examined to explore how environmental factors impact RF fingerprinting, such as transmitter locations, transmitter distance, and device configurations. We identify the most challenging scenarios and study how environmental factors lead to model deterioration through t-SNE visualization.

Modeling wireless channels accurately remains a challenge due to environmental variations and signal uncertainties. Recent neural networks can learn radio frequency (RF) signal propagation patterns, but they process each voxel on the ray independently, without considering global context or environmental factors. Our paper presents a new approach that learns comprehensive representations of complete rays rather than individual points, capturing more detailed environmental features. We integrate a Kolmogorov-Arnold network (KAN) architecture with transformer modules to achieve better performance across realistic and synthetic scenes while maintaining computational efficiency. Our experimental results show that this approach outperforms existing methods in various scenarios. Ablation studies confirm that each component of our model contributes to its effectiveness. Additional experiments provide clear explanations for our model’s performance.

Wireless signals are widely used for human sensing, but they require devices and targets to remain stationary, especially for fine-grained motions like respiration. To enable vital sign monitoring under motion using RFID, we employ dual tags to create a relative coordinate system that reduces motion interference. We also propose physics-informed generative models with frequency domain constraints to improve noise reduction, capturing both time and frequency features. Our method, tested across dynamic scenarios including walking, treadmill exercises, and driving and validated using real patient data, demonstrates superior performance compared to traditional approaches in accurately matching real respiratory signals and exhibits robustness against time shifts.

Self-supervised learning (SSL) has emerged as a powerful deep learning paradigm for human activity recognition (HAR) systems. By leveraging large amounts of unlabeled data, SSL enables the development of pre-trained models (PTMs) that can be efficiently fine-tuned with limited labeled data for downstream HAR tasks, reducing labeling costs while improving generalization. Despite these advantages, the potential vulnerabilities of SSL-based PTMs in IoT sensing systems have not been sufficiently explored. This paper investigates data-free backdoor attacks on these PTMs, focusing on a practical scenario where attackers cannot access downstream task data. To realize these attacks, we design a set of triggers and predefined output representations (PORs). By mapping triggers to PORs through backdoor training, we can implant backdoor behaviors into the PTMs, thereby introducing vulnerabilities across different downstream sensing tasks without requiring prior knowledge. Extensive experiments show our attack applies broadly across various sensing modalities, data, and PTM architectures.

Deep learning has revolutionized the use of inertial measurement unit (IMU) sensors in mobile applications, such as human activity recognition. Building on the success of pre-trained models across various domains, recent studies have increasingly adopted self-supervised learning (SSL) for a range of sensing tasks. While these SSL approaches improve generalization and reduce labeling requirements, their privacy implications have received limited attention. This paper addresses this gap by examining IMU data privacy during pre-training through membership inference. Our work serves two important purposes: First, it enables data owners to verify if their data was used without permission in encoder pre-training. Second, it demonstrates how adversaries might compromise sensitive human sensing data used in pre-training. To enhance the practicality of membership inference on unlabeled IMU sensing data across different SSL algorithms, we introduce an activity labeling module and a novel perturbation strategy to exploit encoder overfitting characteristics on training data. When an encoder over-fits, it memorizes training data rather than learning generalizable patterns. Therefore, when comparing the original data to the perturbed version, the encoder generates more distinct feature vectors for samples from its training set than for samples it has never seen before. We evaluate our membership inference methods on two mainstream SSL methods across multiple datasets, demonstrating that our method can achieve relatively high precision and recall at low false positive rates.

In communication network management, prediction of mobile network traffic is essential to ensure efficient system operation. Although significant progress has been made in the application of neural networks to traffic prediction tasks, traditional models still face considerable challenges when handling high-dimensional and highly time-dependent data. To address these issues, this paper proposes a new prediction framework that leverages large language models (LLMs), by constructing efficient prompts to enhance the ability of large language models (LLMs) in traffic prediction and improve their understanding of complex traffic patterns. Specifically, we introduce functional data analysis (FDA), a technique that offers superior capabilities compared to traditional methods in processing continuous and high-dimensional data structures, to preprocess traffic data and extract key features. Extensive experiments conducted on multiple LLMs using a real-world dataset validate the effectiveness and scalability of the proposed method. The experimental results demonstrate that the framework achieves significant improvements in predictive performance, providing a promising and efficient solution for traffic data analysis in future communication networks.

Deep learning-based Wi-Fi sensing has received massive interest thanks to the prevalence of Wi-Fi technology. While deep learning techniques provide promising results in Wi-Fi sensing, there are only very few studies on the vulnerabilities against Wi-Fi ensing. In this paper, we studied evasion attacks against deep learning-based Wi-Fi sensing and the countermeasure and conducted an extensive experimental evaluation using two publicly available datasets, namely SignFi and Widar. Accordingly, we proposed three white-box and two black-box attacks and revealed that even with an undetectable power change, evasion attacks can achieve a remarkable attack success rate (ASR) of 97.0% and 95.6% in white-box and black-box settings, respectively. These results highlight the urgent need for countermeasures against evasion attacks in Wi-Fi sensing systems. We introduced adversarial training and randomised smoothing, which notably improved the robustness of the Wi-Fi sensing model. The ASRs for white-box and black-box attacks were reduced to a minimum of around 6% and 2%, respectively. Moreover, randomised smoothing also introduced certifiable robustness, achieving 70.1% of samples certified for our model. The certification method provides an additional layer of reliability, ensuring that the model's performance remains consistent and predictable even under adversarial conditions.

Despite the proven capabilities of deep neural networks (DNNs) in identifying devices through radio frequency (RF) fingerprinting, the security vulnerabilities of these deep learning models have been largely overlooked. While the threat of backdoor attacks is well-studied in the image domain, few works have explored this threat in the context of RF signals. In this paper, we thoroughly analyze the susceptibility of DNN-based RF fingerprinting to backdoor attacks, focusing on a more practical scenario where attackers lack access to control model gradients and training processes. We propose leveraging explainable machine learning techniques and autoencoders to guide the selection of trigger positions and values, allowing for the creation of effective backdoor triggers in a model-agnostic manner. To comprehensively evaluate this backdoor attack, we employ four diverse datasets with two protocols (Wi-Fi and LoRa) across various DNN architectures. Given that RF signals are often transformed into the frequency or time-frequency domains, this study also assesses attack efficacy in the time-frequency domain. Furthermore, we experiment with potential detection and defense methods, demonstrating the difficulty of fully safeguarding against our proposed backdoor attack. Additionally, we consider the attack performance in the domain shift case.

Federated learning (FL), an emerging distributed machine learning paradigm, has been applied to various privacy-preserving scenarios. However, due to its distributed nature, FL faces two key issues: the non-independent and identical distribution (non-IID) of user data and vulnerability to Byzantine threats. To address these challenges, in this paper, we propose FedCAP, a robust FL framework against both data heterogeneity and Byzantine attacks. The core of FedCAP is a model update calibration mechanism to help a server capture the differences in the direction and magnitude of model updates among clients. Furthermore, we design a customized model aggregation rule that facilitates collaborative training among similar clients while accelerating the model deterioration of malicious clients. With a Euclidean norm-based anomaly detection mechanism, the server can quickly identify and permanently remove malicious clients. Moreover, the impact of data heterogeneity and Byzantine attacks can be further mitigated through personalization on the client side. We conduct extensive experiments, comparing multiple state-of-the-art baselines, to demonstrate that FedCAP performs well in several non-IID settings and shows strong robustness under a series of poisoning attacks.

Human activity recognition (HAR) based on wearable devices has become an active research direction in the field of ubiquitous computing, and has a wide range of Internet of Things (IoT) applications. Unfortunately, it is challenging to obtain large amounts of labeled sensing data, and manual annotation is time-consuming and labor-intensive, making it impossible for the extensive deployment of HAR systems. Consequently, self-supervised learning has emerged to address this challenge by training on unlabeled data. However, traditional contrastive learning fails to simulate more sample diversity problems caused by environmental heterogeneity and sensor heterogeneity. In this paper, we propose a multi-positive sample quantum contrastive learning (MPSQCL) framework. By increasing the positive samples for contrastive learning and leveraging the advantages of quantum machine learning (QML) techniques, the richer features of input samples are extracted to improve the robustness and generalization of the model. Moreover, we design a new contrastive loss function to adapt to multiple positive sample contrastive learning scenarios. Finally, we validate the effectiveness of the proposed framework on several publicly available HAR datasets.

Recently, there has been a surge in cyberattacks targeting the Internet of Health Things (IoHT), increasing the urgency for advancing network intrusion detection systems (IDS). Machine learning techniques, especially deep neural networks (DNNs), are demonstrating potential in improving the precision of detection methods. Despite their advantages, the complexity of DNNs can obscure their decision-making process, impacting their acceptance in security-critical environments. To improve the transparency of DNN models, we propose a novel post-hoc interpretation method that applies a perturbation-based approach with an optimized mask applied to an autoencoder model for IDS. More importantly, we leverage contrastive learning to maintain perturbed samples within the original feature space, reducing the risk of misclassification due to sample drift and ensuring a clear interpretation of the mask. We validate our approach using the NSL-KDD and UNSW15 datasets, showing that it provides clearer and more robust explanations compared to existing methods. This enhancement in interpretability is pivotal for healthcare cybersecurity experts to gain insights into the decision-making processes of black-box models.

Recent years have witnessed rapid development of Wi-Fi sensing applications. However, the domain shift problem is still an open problem. Variations in environment, time, and detected objects can undermine the effectiveness of cross-domain sensing. This paper proposes a few-shot learning framework for Wi-Fi sensing that enables generalization to unseen domains given only a few samples. To better extract stable features, functional data analysis (FDA) is first employed as a preprocessing technique. We thoroughly evaluate our approach to different Wi-Fi sensing tasks: gesture recognition, and activity recognition. Our experimental results demonstrate that FDA assisted system improves cross-domain accuracy by 14%, 10%, and 8% on the respective tasks with five samples per class.

Heartbeat signals are useful to disease prediction, sub-health diagnosis, fatigue warning, and even emotion estimation. There is a compelling need for contactless, easy-to-deploy, and long-term heartbeat monitoring. This paper presents a contactless Radio Frequency Identification (RFID) based system for heartbeat monitoring that leverages the insight that RFID signal fluctuations induced by chest motion are synchronous with both respiration and heartbeat. The proposed system collects the temporal phase information from the tag pair on the body to extract heartbeat signals using a sequence of signal processing techniques. We propose a signal separation method based on empirical mode decomposition (EMD) to obtain heart rate after preprocessing. Furthermore, the estimated signal is input to an enhanced variational autoencoder (VAE) model to recover the heartbeat waveform. Implemented with commercial off-the-shelf (COTS) RFID devices, the system achieves accurate heart rate monitoring with less than 3% relative errors. The detected waveform exhibits a median cosine similarity of 0.83 as compared with the ground truth, which validate the system’s wide applicability and high reliability for fine-grained, contactless heartbeat monitoring.

Large language models (LLMs) have gained considerable attention for their excellent natural language processing capabilities. Nonetheless, these LLMs present many challenges, particularly in the realm of trustworthiness. This paper introduces TrustLLM, a comprehensive study of trustworthiness in LLMs, including principles for different dimensions of trustworthiness, established benchmark, evaluation, and analysis of trustworthiness for mainstream LLMs, and discussion of open challenges and future directions. Specifically, we first propose a set of principles for trustworthy LLMs that span eight different dimensions. Based on these principles, we further establish a benchmark across six dimensions including truthfulness, safety, fairness, robustness, privacy, and machine ethics. We then present a study evaluating 16 mainstream LLMs in TrustLLM, consisting of over 30 datasets. Our findings firstly show that in general trustworthiness and capability (i.e., functional effectiveness) are positively related. Secondly, our observations reveal that proprietary LLMs generally outperform most open-source counterparts in terms of trustworthiness, raising concerns about the potential risks of widely accessible open-source LLMs. However, a few open-source LLMs come very close to proprietary ones, suggesting that open-source models can achieve high levels of trustworthiness without additional mechanisms like moderator, offering valuable insights for developers in this field. Thirdly, it is important to note that some LLMs may be overly calibrated towards exhibiting trustworthiness, to the extent that they compromise their utility by mistakenly treating benign prompts as harmful and consequently not responding. Besides these observations, we’ve uncovered key insights into the multifaceted trustworthiness in LLMs. We emphasize the importance of ensuring transparency not only in the models themselves but also in the technologies that underpin trustworthiness. We advocate that the establishment of an AI alliance between industry, academia, the open-source community to foster collaboration is imperative to advance the trustworthiness of LLMs.

In this paper, we propose a cross-domain, scalable, and interpretable radio frequency (RF) fingerprinting system using a modified prototypical network (PTN) and an explanation-guided data augmentation across various domains and datasets with only a few samples. Specifically, a convolutional neural network is employed as the feature extractor of the PTN to extract RF fingerprint features. The predictions are made by comparing the similarity between prototypes and feature embedding vectors. To further improve the system performance, we design a customized loss function and deploy an eXplainable Artificial Intelligence (XAI) method to guide data augmentation during fine-tuning. To evaluate the effectiveness of our system in addressing domain shift and scalability problems, we conducted extensive experiments in both cross-domain and novel-device scenarios. Our study shows that our approach achieves exceptional performance in the cross-domain case, exhibiting an accuracy improvement of approximately 80% compared to convolutional neural networks in the best case. Furthermore, our approach demonstrates promising results in the novel-device case across different datasets. Our customized loss function and XAI-guided data augmentation can further improve authentication accuracy to a certain degree.

Despite the proven capabilities of deep neural networks (DNNs) for radio frequency (RF) fingerprinting, their security vulnerabilities have been largely overlooked. Unlike the extensively studied image domain, few works have explored the threat of backdoor attacks on RF signals. In this paper, we analyze the susceptibility of DNN-based RF fingerprinting to backdoor attacks, focusing on a more practical scenario where attackers lack access to control model gradients and training processes. We propose leveraging explainable machine learning techniques and autoencoders to guide the selection of positions and values, enabling the creation of effective backdoor triggers in a model-agnostic manner. To comprehensively evaluate our backdoor attack, we employ four diverse datasets with two protocols (Wi-Fi and LoRa) across various DNN architectures. Given that RF signals are often transformed into the frequency or time-frequency domains, this study also assesses attack efficacy in the time-frequency domain. Furthermore, we experiment with potential defenses, demonstrating the difficulty of fully safeguarding against our attacks.

In this paper, we propose a novel approach to cross-domain unmanned aerial vehicle (UAV) authentication using radio frequency (RF) fingerprinting based on prototypical networks (PTNs). UAVs present a unique challenge for RF fingerprinting due to their hovering motion, which creates more diverse signal domains compared to other RF devices like Wi-Fi. This results in a severe domain shift problem, where well-trained models struggle to generalize to unseen domains. To address this issue without incurring significant costs in data collection and model retraining, we employ PTNs, a few-shot learning paradigm that enhances cross-domain performance and system viability. We further improve our method's effectiveness by incorporating fine-tuning with data augmentation, maintaining system viability while improving performance. Comprehensive experimental results demonstrate that our approach significantly mitigates domain shift, achieving up to a 20% improvement in cross-domain accuracy for UAV fingerprinting.

Radio frequency fingerprint identification (RFFI) is a promising physical layer authentication technique that utilizes the unique impairments within the analog front-end of transmitters as distinct identifiers. State-of-the-art RFFI systems are frequently powered by deep learning, which requires extensive training data to ensure satisfactory performance. However, current RFFI studies suffer from a severe lack of training data, which poses challenges in achieving high identification accuracy. In this paper, we propose a federated RFFI system that is particularly suitable for Internet of Things (IoT) networks, which holds a high potential to address the data scarcity challenge in RFFI development. Specifically, all the receivers in an IoT network can pre-train a deep learning-driven feature extractor in a federated and unsupervised manner. Subsequently, a new client can perform fine-tuning on the basis of the pre-trained feature extractor to activate its RFFI functionality. Extensive experimental evaluation was carried out, involving 60 commercial off-the-shelf (COTS) LoRa transmitters and six software-defined radio (SDR) receivers. The experimental results demonstrate that the federated RFFI protocol can effectively improve the identification accuracy from 63% to 95%, and is robust to receiver hardware and location variations.

Indoor localization availability is still sporadic in industry, especially at the black-hole, i.e., there only exist cellular signals, no GPS or WiFi signals. Based on our 2-year observations at the DiDi ride-hailing platform in China, there are$ 68\,\text{k}$orders everyday created at black-hole. In this paper, we presentTransparentLoc, a large-scale cellular localization system for pickup position recommendation of the DiDi platform. Specifically, we design a CNN model for real-time localization based on a crowdsourcing fingerprint set constructed by outdoor trajectories and abnormal cell tower detection. Then we leverage a DeepFM model to recommend an optimal pickup position for passengers. We share our 2-year experience with 50 million orders across 13 million devices in 4541 cities to address practical challenges including sparse cell towers, unbalanced user fingerprints, temporal variations, and abnormal cell towers in terms of four major service metrics, i.e., pickup position error, over-30-meters ratio, cancel ratio, and call ratio. The large-scale evaluations show that our system achieves a$ 0.54\,\text{m}$lower median pickup position error compared to the iOS built-in cellular localization system, regardless of environmental changes, smartphone brands/models, time, and cellular providers. Additionally, the over-30-meters ratio, cancel ratio, and call ratio have significant reductions of 0.88%, 0.88%, and 5.13%, respectively.

A skyrocketing increase in cyber-attacks significantly elevates the importance of secure software development. Companies launch various bug-bounty programs to reward ethical hackers for identifying potential vulnerabilities in their systems before malicious hackers can exploit them. One of the most difficult decisions in bug-bounty programs is appropriately rewarding ethical hackers. This paper develops a model of an inter-temporal reward strategy with endogenous e-hacker behaviors. We formulate a novel game model to characterize the interactions between a software vendor and multiple heterogeneous ethical hackers. The optimal levels of rewards are discussed under different reward strategies. The impacts of ethical hackers’ strategic bug-hoarding and their competitive and collaborative behaviors on the performance of the program are also evaluated. We demonstrate the effectiveness of the inter-temporal reward mechanism in attracting ethical hackers and encouraging early bug reports. Our results indicate that ignoring the ethical hackers’ strategic behaviors could result in setting inappropriate rewards, which may inadvertently encourage them to hoard bugs for higher rewards. In addition, a more skilled e-hacker is more likely to delay their reporting and less motivated to work collaboratively with other e-hackers. Moreover, the vendor gains more from e-hacker collaboration when it could significantly increase the speed or probability of uncovering difficult-to-detect vulnerabilities.

Introduction to the Special Section on Contact-free Smart Sensing in AloTArtificial Intelligence (AI) and the Internet of Things (IoT) are two powerful forces that have been reshaping our world in recent years.When they converge, they create a new field of AIoT that enables ubiquitous intelligence through the integration of smart algorithms and connected devices.One of the key enablers of AIoT is contact-free sensing, which leverages the availability of portable and highly integrated WiFi, radar, and sonar-style devices to monitor humans and environments without physical contact.This technology has transformed the traditional computer vision-based paradigms and opened up novel possibilities for data collection and analysis.However, contactfree sensing also poses new challenges and risks for AIoT applications.The dynamic and complex wireless environments require innovative solutions for efficient data processing and interpretation.The security and privacy issues of WiFi, radar, and sonar-enabled sensing devices also demand urgent attention, as they may expose sensitive information to malicious attacks.Therefore, it is imperative to explore the potential and pitfalls of contact-free sensing in AIoT and to develop effective strategies for ensuring the robustness and reliability of AIoT applications.This special issue is dedicated to highlighting the cutting-edge methods and latest research in the field of contact-free sensing, which leverages WiFi, radar, and sonar-style devices to monitor humans and environments without physical contact.The main focus of this issue is to explore the latest machine learning analytics to extract information from the sensory data and to investigate the potential risks and countermeasures to ensure the security and privacy of sensing devices.The call for papers attracted with 44 submissions and after a rigorous review, 18 papers have been accepted for this special issue.A brief summary of some papers in this special issue is presented in the following:In "Feasibility of Remote Blood Pressure Estimation via Narrow-band Multi-wavelength Pulse Transit Time, " the authors investigate the feasibility of estimating blood pressure (BP) via pulse transit time (PTT) in a novel remote single-site manner using a modified RGB camera.A narrowband triple band-pass filter makes it possible to measure the PTT between different skin layers, harvesting information from green and near-infrared wavelengths.They design a color-channel model and a novel channel-separation method to further resolve the inter-channel influence and band overlap.The results showed a good absolute Pearson's correlation coefficient between both MW PTT and systolic BP as well as diastolic BP, pointing to the feasibility of the proposed novel remote MW BP estimation via PTT.In "LiteWiSys: A Lightweight System for WiFi-based Dual-task Action Perception, " Sheng et al. propose a lightweight system named LiteWiSys that can simultaneously detect and recognize WiFi-based human actions.This work addresses two major drawbacks of existing methods: heavy

Recently, wireless sensing techniques have been widely used for Internet of Things (IoT) applications. Unlike traditional device-based sensing, wireless sensing is contactless, pervasive, low cost, and non-invasive, making it highly suitable for relevant IoT applications. However, most existing methods are highly dependent on high-quality datasets, and the minority class will not achieve a satisfactory performance when suffering from a class imbalance problem. In this article, we propose a time–frequency semantic generative adversarial network framework (i.e., TFSemantic) to address the imbalanced classification problem in human activity recognition using radio frequency (RF) signals. Specifically, the TFSemantic framework can learn semantic features from the minority classes and then generate high-quality signals to restore data balance. It includes a data pre-processing module, a semantic extraction module, a semantic distribution module, and a data augmenter module. In the data pre-processing module, we process four different RF datasets (i.e., WiFi, RFID, UWB, and mmWave). We also develop Fourier semantic feature convolution and attention semantic feature embedding methods for the semantic extraction module. A discrete wavelet transform is utilized for reconstructed RF samples in the semantic distribution module. In data augmenter module, we design an associated loss function to achieve effective adversarial training. Finally, we validate the effectiveness of the proposed TFSemantic framework using different RF datasets, which outperforms several state-of-the-art methods.

To implement ubiquitous wireless sensing, the domain shift problem (e.g., different environments, users, devices) for machine learning based approaches should be addressed. Some existing methods are proven to be effective, such as transfer learning and domain adaptation. Meanwhile, quantum machine learning, a combination of quantum computing and machine learning, has attracted much attention. More importantly, quantum transfer learning (QTL) has been successful for certain applications, e.g., image classification. In this paper, we explore a classical to quantum (C2Q) framework to address the domain shift problem in wireless sensing by exploiting the great potential of QTL. Specifically, we first analyze the data shift problem in various types of wireless datasets by calculating the Kullback-Leibler (KL) divergence of different domains. Then, a QTL framework is designed to introduce importance weighting and adversarial strategies. We finally evaluate the proposed framework using the representative human activity recognition task on three wireless sensing datasets. Experimental results demonstrate the feasibility of the framework and its great potential for solving the domain shift problem in wireless sensing.

Millimeter wave (mmWave) communications and massive MIMO play crucial roles in the development of future wireless systems. In addition to offering high data rates, these technologies enable the realization of high-precision localization systems, especially in complicated indoor rich multi-path environments without GPS coverage. While deep neural networks (DNNs) enable high accuracy in fingerprint-based indoor localization, their implementations also introduce security problems. In the field of computer vision, backdoor attacks have proven to be able to effectively deceive models using specific or imperceptible triggers. In this paper, we study the impact of backdoor attacks on 5G massive MIMO localization systems in both indoor and outdoor environments. Two different triggers are investigated: the one-pixel trigger (visible) and the random noise trigger (invisible). We evaluate the localization systems using a public dataset and demonstrate that DNN-based localization systems are vulnerable to backdoor attacks.

Location awareness is the basis for enabling pickup service at ride-hailing platforms. In contrast to the almost pervasive coverage outdoors, indoor localization availability is still sporadic in industry since it largely relies on RF signatures from certain IT infrastructure, e.g., WiFi access points. Based on our 2-year observations at DiDi ride-hailing platform in China, there are 68k orders everyday created at black-hole, i.e., where only cellular signals exist. In this paper, we present the design, development, and deployment of TransparentLoc, a large-scale cellular localization system for pickup position recommendation, and share our 2-year experience with 50 million orders across 13 million devices in 4541 cities to address practical challenges including sparse cell towers, unbalanced user fingerprints, and temporal variations. Our system outperforms the iOS built-in cellular localization system in terms of four major service metrics, regardless of environmental changes, smartphone brands/models, time, and cellular providers.

With the rapid development towards smart Internet of Things (IoT), detection of human activity has become essential in a variety of applications. Various radio-frequency (RF) sensing technologies, such as WiFi, Radio-Frequency Identification (RFID), and Frequency-Modulated Continuous Wave (FMCW) radar, have been utilized for non-invasive human activity recognition (HAR). It will be highly desirable to develop a HAR solution that can work with different types of RF technologies, such that the cost and the barrier of wide deployment can both be greatly reduced, and more robust performance can be achieved by utilizing the complementary RF sensory data. In this paper, we propose a technology-agnostic approach for RF-based HAR, termed TARF, which works with several different RF sensing technologies. A novel data generalization technique is proposed to mitigate the disparity in measured data from different RF devices. A domain adversarial neural network is proposed to combat the interference from various RF sensing technologies. The performance of the proposed system is evaluated with experiments using four different RF sensing technologies. TARF is shown to outperform the state-of-the-art Convolutional Neural Network (CNN)-based solution with considerable gains.

With the wide deployment of 5G communication systems, 5G massive multiple-input multiple-output (MIMO) has been shown effective not only to improve the spectrum efficiency and energy efficiency, but also provides location-based service (LBS) such as outdoor vehicle localization and indoor user localization. Recently, deep convolutional neural network (DCNN) has been applied for massive MIMO localization using channel state information (CSI) or angle-delay profile (ADP). However, the robustness of the DCNN model has not been explored in massive MIMO localization. In this paper, we study the impact of adversarial attack and defense (i.e., adversarial training) on massive MIMO localization using DCNN and the neural ordinary differential equation (ODE) model. We first introduce the massive MIMO system with respect to the channel model and ADP fingerprints, and then present the DCNN model and the neural ODE model for massive MIMO localization, as well as three types of white-box adversarial attacks and adversarial training. Finally, our experimental results validate that the proposed neural ODE with adversarial training could effectively improve the robustness of massive MIMO localization in indoor and outdoor environments.

Driven by the fast increase of food demand world wide, the safety of grain storage becomes increasingly important. TWo key factors, i.e., temperature and moisture, greatly influence the safety of stored grain. The traditional methods of detecting grain temperature and moisture are time-consuming, expensive, and inconvenient to use. In this paper, we develop a TagSense system for robust wheat moisture and temperature sensing using cheap commercial-off-the-shelf (COTS) RFID devices. We first validate the feasibility of using tag impedance for robust moisture and temperature sensing. We then propose a distance- free algorithm and an angle-agnostic method to mitigate the impact of different measurement distances and angles. Our experimental results demonstrate that the TagSense system can achieve satisfactory sensing accuracy of wheat moisture and temperature in different rotation angles and at different distances.

Radio frequency (RF) fingerprinting is a hardware feature used in Internet of Things (IoT) applications to identify wireless devices. In this paper, we propose few-shot learning (FSL) and prototypical networks (PTNs) to create a new model that can adapt to a new domain with very few labeled examples. The proposed model can mitigate the domain shift caused by changing RF environments. Experimental results show the proposed method can improve the performance of RF fingerprinting over different domains.

Fingerprinting-based indoor localization has been a research focus for GPS denied areas. The development of neural networks has greatly promoted its application in indoor localization systems. However, recent studies showed that the machine learning models, including state-of-the-art neural networks, are vulnerable to adversarial examples, and thus, neural network-based indoor localization systems are also under the threat of adversarial attacks. To investigate the effect of adversarial attacks on indoor localization systems and to make such systems resilient to adversarial attacks, we propose AdvLoc, an adversarial deep learning for indoor localization system. With the proposed AdvLoc system, the effect of adversarial attacks on indoor localization is studied under six types of adversarial attack methods in both black-box attack and white-box attack scenarios. Furthermore, adversarial training is utilized in offline training of the proposed AdvLoc system, which is effective against first-order adversarial attacks. The proposed AdvLoc system is implemented with commodity WiFi devices and evaluated with extensive experiments in two representative indoor environments. The experimental results verify the robustness of the proposed system against first-order adversarial attacks in representative indoor environments.

With the rapid growth of the Internet of Things (IoT), the radio-frequency identification (RFID) technology has been recognized as an effective and low-cost solution for many IoT applications. In this article, we study the problem of utilizing a sparse RFID tag array for backscatter indoor localization. We first theoretically and experimentally validate the feasibility of using sparse tag arrays for the direction of arrival (DOA) estimation. We then present the SparseTag system, which leverages a novel sparse tag array for high-precision backscatter indoor localization. The SparseTag system includes sparse array processing, difference co-array design, DOA estimation using a spatial smoothing-based method, and a localization method. A robust channel selection method based on the RFID tag array is adopted for mitigating the multipath effect. The SparseTag system is implemented with commodity RFID devices. Its superior performance is validated in two different environments with extensive experiments and comparison to baseline schemes.

Human pose tracking has attracted great interest re-cently. Considerable efforts have been made in Radio-Frequency (RF) sensing techniques for human pose tracking without using a video camera. Although the existing RF based schemes can well protect user privacy, they are usually sensitive to the RF environment and are hard to generalize to new environments. In this paper, we analyze the challenges of generalization of Radio-Frequency Identification (RFID) based human pose tracking systems. We then present an RFID based 3D human pose tracking system, termed Meta-Pose, which incorporates meta-learning and few-shot fine-tuning to achieve high adaptability to new environments. The proposed system is implemented with commodity RFID devices and extensive experiments are conducted for performance evaluation. The experiment results validate the superior human pose tracking performance and high adaptability of the proposed Meta-Pose system.

Millimeter Wave (mmWave) communications, as a core technique of 5G, can be leveraged for outdoor localization because of its large bandwidth and massive antenna array. Fingerprinting based mmWave outdoor localization methods using deep learning are highly suitable for non-line-of-sight (NLOS) environments. In this paper, we propose a deep convolutional Gaussian process (DCGP) based regression approach to achieve high robustness for fingerprinting-based mmWave outdoor localization, which exploits the convolutional structure for deep Gaussian process to allow uncertainty estimation on location predictions. Specially, we present a system architecture of mmWave based outdoor localization, including beamforming image construction and DCGP training, where DCGP model can effectively learn the location features from mmWave beamforming images. Our experimental results show that the proposed DCGP method can achieve higher outdoor localization accuracy than a CNN-based baseline method.

Vital sign (e.g., respiration rate) monitoring has become increasingly more important because it offers useful clues about medical conditions such as sleep disorders. There is a compelling need for technologies that enable contact-free and easy deployment of vital sign monitoring over an extended period of time for healthcare. In this article, we present a SonarBeat system to leverage a phase-based active sonar to monitor respiration rates with smartphones. We provide a sonar phase analysis and discuss the technical challenges for respiration rate estimation utilizing an inaudible sound signal. Moreover, we design and implement the SonarBeat system, with components including signal generation, data extraction, received signal preprocessing, and breathing rate estimation with Android smartphones. Our extensive experimental results validate the superior performance of SonarBeat in different indoor environment settings.

The development of Internet-of-Things (IoT) technology promotes the advances of grain condition detection and analysis systems. Temperature monitoring is a main element to maintain grain quality, and effective control of grain temperature is crucial to safe storage of grain. In this article, an encoder–decoder model with attention mechanism is proposed to accurately forecast the temperature of stored grain. Considering that the points on the gradient direction of the temperature surface have a great influence on the temperature of the target point, the Sobel operator is used to extract the local characteristics of the target point. In addition, considering the correlation structure in the sensory data, the attention mechanism is used to extract the global features of the target point. The extracted spatial features are fed into long short-term memory (LSTM) networks to obtain the long-term state information of spatial factors. LSTM unit and convolutional neural network are used to encode the spatial features of the target points. Taking meteorological factors as the external input of the decoder, temporal attention mechanism and LSTM unit are used to complete the decoding process and realize the prediction of grain temperature in the future. The results with real grain storage data show that the proposed model outperforms several schemes, including Kalman-modified the least absolute shrinkage and selection operator (Kalman-modified LASSO), temporal graph convolutional network (T-GCN), LSTM, CNN-LSTM, and convolutional LSTM (Conv-LSTM), with considerable gains.

Wi-Fi-based indoor fingerprinting is attracting increasing interest in the research community due to the ubiquitous access in indoor environments. In this article, we propose ResLoc, a deep residual sharing learning-based system for indoor fingerprinting using bimodal channel state information (CSI) tensor data. The proposed ResLoc system employs CSI tensor data, including the angle of arrival and amplitude, collected from a small set of training locations with known coordinates to train the proposed dual-channel deep residual sharing learning model. The proposed new model extends the traditional deep residual learning model by incorporating two or more channels and let the channels exchange their residual signals after each residual block. Unlike prior deep-learning-based fingerprinting schemes, ResLoc only requires for training one group of weights for all the training locations. The proposed ResLoc system is implemented with commodity Wi-Fi devices and evaluated with extensive experiments in three representative indoor environments. The experimental results validate that the proposed ResLoc system can achieve high localization accuracy using a single Wi-Fi access point in indoor environments.

To improve driving safety and avoid accidents caused by driving fatigue, drowsiness detection aims to alarm the driver before he/she falls asleep. Since breathing rate is a key indicator of the drowsy state, respiration monitoring in the noisy driving environment is critical for developing an effective driving fatigue detection system. In this paper, we propose, for the first time, an RFID based respiration monitoring system for driving environments. The system estimates the respiration rate of a driver based on phase values sampled from multiple RFID tags attached to the seat belt, while exploiting the tag diversity to combat the strong noise in the driving environment. Both tensor completion and tensor Canonical Polyadic Decomposition (CPD) are applied to process the phase values, to overcome the influence of frequency hopping, random sampling, vehicle vibration, and other environmental movements. The proposed system is analyzed and implemented with commodity RFID devices. Its accurate and robust performance is demonstrated with extensive experiments conducted in a real driving car.

In recent years, human pose tracking has become an important topic in computer vision (CV). To improve the privacy of human pose tracking, there is considerable interest in techniques without using a video camera. To this end, radio-frequency identification (RFID) tags, as a low-cost wearable sensor, provide an effective solution for 3-D human pose tracking. In this article, we propose RFID-Pose, a vision-aided realtime 3-D human pose estimation system, which is based on deep learning assisted by CV. The RFID phase data are calibrated to effectively mitigate the severe phase distortion, and high accuracy low rank tensor completion is employed to impute the missing RFID data. The system then estimates the spatial rotation angle of each human limb, and utilizes the rotation angles to reconstruct human pose in realtime with the forward kinematic technique. A prototype is developed with commodity RFID devices. High pose estimation accuracy and realtime operation of RFID-Pose are demonstrated in our experiments using Kinect 2.0 as a benchmark.

This paper aims at predicting accurate outdoor and indoor locations using deep neural networks, for the data collected using the Long-Range Wide-Area Network (LoRaWAN) communication protocol. First, we propose an interpolation aided fingerprinting-based localization system architecture. We propose a deep autoencoder method to effectively deal with the large number of missing samples/outliers caused by the large size and wide coverage of LoRa networks. We also leverage three different deep learning models, i.e., the Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), and the Convolutional Neural Network (CNN), for fingerprinting based location regression. The superior localization performance of the proposed system is validated by our experimental study using a publicly available outdoor dataset and an indoor LoRa testbed.

Temperature is one of the major ecological factors that affect the safe storage of grain. In this paper, we propose a deep spatio-temporal attention mode to predict stored grain temperature, which exploits the historical temperature data of stored grain and the meteorological data of the region. In this proposed model, we use the Sobel operator to extract the local spatial factors, and leverage the attention mechanism to obtain the global spatial factors of grain temperature data and temporal information. In addition, a convolutional neural network (CNN) is used to learn features of external meteorological factors. Finally, the spatial factors of grain pile and external meteorological factors are combined to predict future grain temperature using long short-term memory (LSTM) based encoder and decoder models. Experiment results show that the proposed model achieves higher predication accuracy compared with the traditional methods.

With the rapid development of computer vision, human pose tracking has attracted increasing attention in recent years. To address the privacy concerns, it is desirable to develop techniques without using a video camera. To this end, RFID tags can be used as a low-cost wearable sensor to provide an effective solution for 3D human pose tracking. User adaptability is another big challenge in RF based pose tracking, i.e., how to use a well-trained model for untrained subjects. In this paper, we propose Cycle-Pose, a subject-adaptive realtime 3D human pose estimation system, which is based on deep learning and assisted by computer vision for model training. In Cycle-Pose, RFID phase data is calibrated to effectively mitigate the severe phase distortion, and High Accuracy LowRank Tensor Completion (HaLRTC) is employed to impute missing RFID data. A cycle kinematic network is proposed to remove the restriction on paired RFID and vision data for model training. The resulting system is subject-adaptive, achieved by learning to transform the RFID data into a human skeleton for different subjects. A prototype system is developed with commodity RFID tags/devices and evaluated with experiments. Compared with a traditional system RFIDPose, higher pose estimation accuracy and subject adaptability are demonstrated by Cycle-Pose in our experiments using Kinect 2.0 data as ground truth.

Radio Frequency (RF) based human pose estimation techniques have been proposed to generate human pose without using a camera, so people will no longer worry about their privacy. Compared with other RF sensing based systems, Radio Frequency Identification (RFID) provides a promising solution for RF based human pose estimation. RFID tags can be used as wearable sensors because of their small size. The interference caused by the multipath effect is much smaller in the RFID system. The cost of RFID systems is also lower than the advanced radar based systems such as FMCW radar. Thus, we propose the RFID-Pose system for tracking the movements of multiple human limbs in realtime [1]. In the proposed system, RFID tags are attached to the target human joints. The movement of the tags are captured by the phase variations in the responses from each tag. The human pose is reconstructed by estimating rotation angles from RFID data and the initial human skeleton. The vision data will not be needed anymore in the testing phase, so the user's privacy can be well protected.

Vital signs, such as respiration and heartbeat, are useful for health monitoring because such signals provide important clues of medical conditions. Effective solutions are needed to provide contact-free, easy deployment, low-cost, and long-term vital sign monitoring. In this article, we present PhaseBeat to exploit channel state information, in particular, phase difference data to monitor breathing and heart rates with commodity WiFi devices. We provide a rigorous analysis of channel state information phase difference with respect to its stability and periodicity. Based on the analysis, we design and implement the PhaseBeat system with off-the-shelf WiFi devices and conduct an extensive experimental study to validate its performance. Our experimental results demonstrate the superior performance of PhaseBeat over existing approaches in various indoor environments.

With the increasing demand for location-based service, WiFi-based localization has become one of the most popular methods due to the wide deployment of WiFi and its low cost. To improve this technology, we propose DeepMap, a deep Gaussian process for indoor radio map construction and location estimation. Received signal strength (RSS) samples are used in DeepMap to generate accurate and fine-grained radio maps. A two-layer deep Gaussian process model is designed to determine the relationship between the location and RSS samples, while the model parameters are optimized with an offline Bayesian training method. To identify the location of a mobile device, a Bayesian fusion method is proposed, which leverages RSS samples from multiple access points (APs) to achieve high location estimation accuracy. We conduct comprehensive experiments to verify the performance of DeepMap in two indoor settings. DeepMap's robustness is validated using limited training data.

In this paper, we propose a real-time, nondestructive, and low-cost wheat mildew detection system using commodity WiFi devices, which is a new application of the Internet of Things (IoT) to agriculture applications. We first introduce wheat mildew and validate the feasibility of wheat mildew detection using WiFi Channel State Information (CSI) amplitude data. We then present the MiFi system design, including CSI sensing, preprocessing, radial basis function (RBF) neural network based detection modeling, and mildew detection. Our experimental results validate the effectiveness of the proposed MiFi system. The average detection accuracy of the MiFi system is over 90% under both line-of-sight (LOS) and non-line-of-sign (NLOS) scenarios.

With the growth of the number of vehicles and car accidents, driving safety is becoming increasingly important. There is a compelling need for an effective, low-cost driving fatigue detection system. In this paper, we propose an RFID based system, termed NodTrack, to detect the nodding movements of drivers, which is a key indicator of fatigue and one of the most dangerous motions during drowsy driving. The NodTrack system utilizes the phase difference between two RFID tags mounted on the back of a hat worn by the driver, to extract nodding features. We propose an effective technique to mitigate the cumulative error caused by frequency hopping in most FCC-compliant RFID systems, as well as a long short-term memory (LSTM) autoencoder model to learn the nodding features from calibrated data. The highly accurate detection performance of the proposed system is validated by our experimental study.

In this paper, we study the problem of utilizing a sparse RFID tag array for backscatter indoor localization. We first theoretically and experimentally validate the feasibility of using RFID tag array for direction of arrival (DOA) estimation. We then present the SparseTag system, which leverages a novel sparse RFID tag array for high-precision backscatter indoor localization. The SparseTag system includes sparse array processing, difference co-array design, DOA estimation using a spatial smoothing based method, and a localization method, while a robust channel selection method based on the RFID tag array is proposed for mitigating the indoor multipath effect. The SparseTag system is implemented with commodity RFID devices. Its superior performance is validated in two different environments with extensive experiments.

In this paper, we present DeepMap, a deep Gaussian process for indoor radio map construction and location estimation. To address the shortcomings of existing Gaussian process based approaches, we present a DeepMap system, which employs deep Gaussian process for constructing received signal strength (RSS) radio maps and a Bayesian algorithm for online localization. We design a two-layer deep Gaussian process model to capture the relationship between the RSS space and the location space and provide an offline Bayesian training method to determine model parameters. A Bayesian fusion method using multiple APs is proposed for accurate location estimation. Experimental results verify the performances of DeepMap in a large indoor environment and validate its robustness with moderate training data.

With the growth of smart healthcare in the Internet of Things (IoT), breathing monitoring and apnea detection are of increasing importance. In this paper, we propose AutoTag, a recurrent variational autoencoder model for breathing and apnea detection with commodity RFID Tags. The AutoTag system consists of signal extraction, calibration, and respiration monitoring modules. We propose a novel method to mitigate the frequency hopping offset with realtime calibration for FCC complaint RFID systems, and a new recurrent variational autoencoder method for apnea and breathing detection. Experimental results demonstrate the effectiveness of the proposed AutoTag system in two different environments.

With the fast increasing demands of location-based service and proliferation of smartphones and other mobile devices, accurate indoor localization has attracted great interest. In this paper, we present DeepML, a deep long short-term memory (LSTM) based system for indoor localization using the smartphone magnetic and light sensors. We verify the feasibility of using bimodal magnetic and light data for indoor localization through experiments. We then design the DeepML system, which first builds bimodal images by data preprocessing, and then trains a deep LSTM network to extract the location features. Newly received magnetic field and light intensity data is then exploited for estimating the location of the mobile device using an improved probabilistic method. Our extensive experiments verify the effectiveness of the proposed DeepML system.

In this paper, we present a non-destructive and economic wheat moisture detection system with commodity WiFi. First, we experimentally validate the feasibility of wheat moisture detection by using CSI amplitude and phase difference data. We then design Wi-Wheat system, where data preprocessing, feature extraction and support vector machine (SVM) classification are implemented for CSI processing module. For data preprocessing, we employ outlier detection, data normalization and eliminating noise for obtaining clear CSI amplitude and phase difference data. Then, we consider principal component analysis (PCA) based feature extraction for Wi-Wheat system. For SVM classification, Gaussian radial basis function (RBF) is used as the kernel function for wheat moisture detection. The experimental results show the Wi-Wheat system can achieve higher classification accuracy for LOS and NLOS scenarios.

Moisture content of cereal grains is a highly important factor in safe storage and food processing. The existing detection methods are either time-consuming, sensitive to the environment, or have a high cost. In this paper, we propose DeepWMD, a deep LSTM network based system for multi-class wheat moisture detection. We first collect CSI amplitude and phase difference data to detect wheat moisture content. Then, we design the DeepWMD system with commodity Wi-Fi devices in the 5GHz band, including data preprocessing of collected CSI data, offline training, and online testing. Our experimental results verify the efficacy of the proposed DeepWMD system, and demonstrates that DeepWDM can achieve high-precision multi-class wheat moisture detection in different indoor storage environments.

With the increasing commercial prospect of personal Unmanned Aerial Vehicle (UAV), human and UAV interaction has been a compelling and challenging task. In this paper, we present the RFHUI, a human and UAV interaction system based on passive radio-frequency identification (RFID) technology which provides a remote control function. Three or more Ultra high frequency (UHF) RFID tags are attached on a board to create a hand-held controller. A COTS (Commercial Off-The-Shelf) RFID reader with multiple antennas is deployed to collect the observations of the tags. According to the phase measurement from the RFID reader, we leverage a Bayesian filter based method to localize the position of all tags in a global coordinate. From the estimated position of the attached tags, a 6 DOF (Degrees of Freedom) pose of the controller can be obtained. Therefore, when the user moves the controller, its pose will be precisely tracked in a real-time manner. Then, the flying commands, which are generated from the estimated pose of the controller, are sent to the UAV for navigation. We implemented a prototype of the RFHUI, and the experiment results show that it provides precise poses with 0.045 m error in position and 2.5° error in orientation for the controller. It therefore enables the controller to precisely and intuitively instruct the UAV's navigation in an indoor environment.

Vital signs, such as breathing rate, can provide useful information for personal healthcare. In this paper, we present ResBeat, a commodity 5GHz WiFi based system to exploit bimodal channel state information (CSI), including amplitude and phase difference, for realtime, long- term, and contact-free breathing monitoring. We first present an analysis of breathing signal anomaly based on bimodal CSI data. We then describe the data preprocessing, adaptive signal selection, and breathing signal monitoring modules of ResBeat, and employ peak detection to estimate breathing rates. We conduct extensive experiments under three different environments, where superior performance over two alternative methods is validated.

With the increasing demand of location-based services, Wi-Fi based localization has attracted great interest because it provides ubiquitous access in indoor environments. In this paper, we propose CiFi, deep convolutional neural networks (DCNN) for indoor localization with commodity 5GHz WiFi. First, by leveraging a modified device driver, we extract phase data of channel state information (CSI), which is used to estimate angle of arriving (AOA). We then create estimated AOA images as input to the DCNN, to train the weights in the offline phase. The location of mobile device is predicted based on the trained DCNN and new CSI AOA images. We implement the proposed CiFi system with commodity Wi-Fi devices in the 5GHz band and verify its performance with extensive experiments in two representative indoor environments.

Vital sign (e.g., breathing rate) monitoring has become increasingly more important because it can offer useful clues to medical conditions such as sleep disorders or anomalies. There is a compelling need for technologies that enable contact-free, easy deployment, and long-term vital sign monitoring for healthcare. In this paper, we present a SonarBeat system to leverage a phase based active sonar to monitor breathing rates with smartphones. We design and implement the SonarBeat system, with components including signal generation, data extraction, received signal preprocessing, and breathing rate estimation, with Andriod smartphones. Our experimental results validate the superior performance of SonarBeat in different indoor environment settings.

Vital signs, such as respiration and heartbeat, are useful to health monitoring since such signals provide important clues of medical conditions. Effective solutions are needed to provide contact-free, easy deployment, low-cost, and long-term vital sign monitoring. In this paper, we present PhaseBeat to exploit channel state information (CSI) phase difference data to monitor breathing and heartbeat with commodity WiFi devices. We provide a rigorous analysis of the CSI phase difference data with respect to its stability and periodicity. Based on the analysis, we design and implement the PhaseBeat system with off-the-shelf WiFi devices, and conduct an extensive experimental study to validate its performance. Our experimental results demonstrate the superior performance of PhaseBeat over existing approaches in various indoor environments.

Wi-Fi based indoor localization has attracted great interest due to its ubiquitous access in many indoor environments. In this paper, we propose ResLoc, a deep residual sharing learning based system for indoor localization with channel state information (CSI) tensor data. We first introduce CSI data in wireless systems and show how to build CSI tensors for indoor localization. Then, we present the design of ResLoc, which employs dual-channel, bi-modal CSI tensor data to train the deep network using the proposed deep residual sharing learning in the offline phase. In the online test phase, we use newly received CSI tensor data to estimate the location of the mobile device based on an enhanced probabilistic method. The experimental results show that the proposed ResLoc system can obtain submeter level accuracy with a single access point.

Vital sign (e.g., breathing rate) monitoring has become increasingly more important because it offers useful clues of medical conditions such as sleep disorders or anomalies. It is necessary to provide contact-free, easy deployment, and long-term vital sign monitoring for healthcare. In this demo, we present SonarBeat to leverage a phase based active sonar to monitor breathing rates with smartphones.

Breathing signal monitoring can provide important clues for health problems. Compared to existing techniques that require wearable devices and special equipment, a more desirable approach is to provide contact-free and long-term breathing rate monitoring by exploiting wireless signals. In this article, we propose TensorBeat, a system to employ channel state information (CSI) phase difference data to intelligently estimate breathing rates for multiple persons with commodity WiFi devices. The main idea is to leverage the tensor decomposition technique to handle the CSI phase difference data. The proposed TensorBeat scheme first obtains CSI phase difference data between pairs of antennas at the WiFi receiver to create CSI tensors. Then canonical polyadic (CP) decomposition is applied to obtain the desired breathing signals. A stable signal matching algorithm is developed to identify the decomposed signal pairs, and a peak detection method is applied to estimate the breathing rates for multiple persons. Our experimental study shows that TensorBeat can achieve high accuracy under different environments for multiperson breathing rate monitoring.

With the increasing demand of location-based services, indoor localization based on fingerprinting has become an increasingly important technique due to its high accuracy and low hardware requirement. In this paper, we propose PhaseFi, a fingerprinting system for indoor localization with calibrated channel state information (CSI) phase information. In PhaseFi, the raw phase information is first extracted from the multiple antennas and multiple subcarriers of the IEEE 802.11n network interface card by accessing the modified device driver. Then a linear transformation is applied to extract the calibrated phase information, which we prove to have a bounded variance. For the offline stage, we design a deep network with three hidden layers to train the calibrated phase data, and employ the weights of the deep network to represent fingerprints. A greedy learning algorithm is incorporated to train the weights layer-by-layer to reduce computational complexity, where a subnetwork between two consecutive layers forms a restricted Boltzmann machine. In the online stage, we use a probabilistic method based on the radial basis function for online location estimation. The proposed PhaseFi scheme is implemented and validated with extensive experiments in two representation indoor environments. It is shown to outperform three benchmark schemes based on CSI or received signal strength in both scenarios.

With the increasing demand of location-based services, indoor localization based on fingerprinting has become an increasingly important technique due to its high accuracy and low hardware requirement. In this paper, we propose PhaseFi, a fingerprinting system for indoor localization with calibrated channel state information (CSI) phase information. In PhaseFi, the raw phase information is first extracted from the multiple antennas and multiple subcarriers of the IEEE 802.11n network interface card (NIC) by accessing the modified driver. Then a linear transform is used to extract the calibrated phase information, which is proven to have a bounded variance. For the offline stage, we design a deep network with three hidden layers to train the calibrated phase data, and employ weights to represent fingerprints. A greedy learning algorithm is incorporated to train the weights layer-by-layer to reduce computational complexity, where a sub-network between two continuous layers forms a Restricted Boltzmann Machine (RBM). In the online stage, we use a probabilistic method based on the radial basis function (RBF) for online location estimation. The proposed PhaseFi scheme is implemented and validated with intensive experiments in two representation indoor environments. It outperforms other three benchmark schemes based on CSI or RSS in both scenarios.

With the fast growing demand of location-based services in indoor environments, indoor positioning based on fingerprinting has attracted a lot of interest due to its high accuracy. In this paper, we present a novel deep learning based indoor fingerprinting system using Channel State Information (CSI), which is termed DeepFi. Based on three hypotheses on CSI, the DeepFi system architecture includes an off-line training phase and an on-line localization phase. In the off-line training phase, deep learning is utilized to train all the weights as fingerprints. Moreover, a greedy learning algorithm is used to train all the weights layer-by-layer to reduce complexity. In the on-line localization phase, we use a probabilistic method based on the radial basis function to obtain the estimated location. Experimental results are presented to confirm that DeepFi can effectively reduce location error compared with three existing methods in two representative indoor environments.

With the proliferation of mobile devices such as smartphones, an interesting problem is how to make use them to improve the accuracy of localization in indoor environments. In this paper, we develop a novel cooperative localization scheme exploiting mobility in the indoor environment. The problem is formulated as a semidefinite program (SDP) using Linear Matrix Inequality (LMI). With the proposed approach, mobile users utilize their top RSS measurements for distance estimation and to mitigate the the shadowing effect found in indoor environments. In addition, we utilize the estimated position for a user from the last time slot as a virtual access point (AP) to obtain the next position estimation, by utilizing the inertial measurement unit (IMU) data from smartphones. To better take advantage of the moving direction and velocity information provided by the smartphones, we next apply Kalman filter to further mitigate the errors in estimated positions. Simulation results confirm that both the mean error and variance can be effectively reduced by exploiting IMU data and Kalman filter.

With the development of wireless sensor network (WSN), the design and maintenance of WSN are still challenge in a large number of sensor nodes which are constrained in energy and bandwidth. Thus, a new solution is high altitude platforms (HAPs) that can be employed as the relay nodes of WSN in order to reduce the energy consumption of sensor nodes because of multi-hop relay transmission and to enlarge the coverage range of WSN. In this paper, a novel heterogeneous WSN system consisting of HAP layer, WSN layer and mission layer is proposed. In the system, we model a deployment model of HAPs based on markov random field with maximum a posteriori probability (MAP) framework, thus obtaining the energy function of HAPs. Then, a potential game approach is introduced to analyze the energy function, which proves to be able to achieve a pure Nash equilibrium. In addition, a modified distributed learning algorithm called sequential spatial adaptive play is presented to solve the proposed potential game. Finally, simulation results illustrate that the proposed method can achieve Nash equilibrium and the optimal deployment of HAPs.

Nowadays, a promising solution to the demand for high capacity wireless services is provided by the high altitude platforms (HAPs), which can exploit the advantages of both satellites and terrestrial networks. However, due to the limited energy of the communication nodes on HAPs, energy-efficient deployment of airships for HAPs becomes increasingly important. In this paper, a heterogeneous network system comprising mission layer, HAP layer and satellite layer is presented. In the system, the mission space is partitioned according to the assumption of one-hop relay which guarantees that all the users are effectively covered by airships in HAP layer. Then, we model an expected energy consumption function of airships, which is minimized using deterministic annealing algorithm. Furthermore, the optimal solution to this approach is analyzed and proved. Simulation results demonstrate that the proposed method can get the goal of the minimum energy consumption and achieve a near optimal deployment of airships.