Intelligent Transportation Systems (ITS) rely on high-precision road networks, which are particularly scarce in indoor parking. Existing methods depend on expensive hardware (e.g., LiDAR) or manual mapping, both of which are costly and inefficient. The rise of the Internet of Things (IoT) has enabled large-scale data collection and connectivity, offering new opportunities for automated road network extraction. OpenStreetMap (OSM), as a crowdsourced IoT-driven platform, provides multi-layer geospatial data, including the Road Network Layer (RNL), Lane Boundary Layer (LBL), and Turn Sign Layer (TSL). However, OSM data often suffers from incompleteness and noisy connectivity, affecting the continuity and accuracy of road networks. This paper introduces OSM2Net, a novel framework designed to extract road networks from individual layers and leverage multi-layer data to construct directed road networks. Specifically, OSM2Net rasterizes noisy OSM data into bitmaps for image processing and multi-layer fusion. By leveraging the topology relationship between lane boundaries and road networks, a Lane-Road Map Generator (LRMG) creates a simulated dataset for training. Then, utilizing the simulated dataset, a Lane2Net model is designed to extract road networks from sparse lane boundary images. The framework then vectorizes bitmaps into a lightweight, undirected road network and refines it into a directed network by extracting and matching turn sign information. Experimental results show that Lane2Net achieves Intersection over Union (IoU) of 93% and 92% using simulated and real-world datasets, respectively. Extensive experiments on real-world datasets confirm that OSM2Net delivers robust completeness and high-quality road network extraction.

Non-line-of-sight (NLOS) identification is the key technique to improve the accuracy of the channel impulse response (CIR) based ultrawideband (UWB) positioning system. However, most existing NLOS identification approaches are tailored to static environments and often encounter difficulties in dynamic settings with both temporal and spatial variations, particularly when dealing with limited and hard samples. This paper introduces a hard sample meta-learning (HSML) approach to address the issues of NLOS identification across different scenarios and domains. HSML includes two phases: a hard sample meta-training phase and a fine-grained meta-testing phase. During the meta-training phase, we train a two-loop learning network using CIR from multiple scenarios (tasks). The inner loop focuses on learning task-specific features, while the outer loop captures cross-task generalization properties using a cross-entropy loss. Hard samples are identified based on estimated residuals for each task, and a new dataset is created, consisting of both hard samples and samples with small residuals. To improve the robustness against hard samples, we implement a residual-corrected focal loss, which is used to retrain the network on this new dataset. In the fine-grained meta-testing phase, we apply a filtering mechanism based on the tendency of estimated residuals during fine-tuning. This mitigates the risk of poor performance caused by anomalous samples. We validate the effectiveness and robustness of the proposed HSML method using two datasets containing multiple real-world scenarios. Our experimental results demonstrate that HSML outperforms existing models in terms of identification accuracy, robustness and generalization performance.

Wireless positioning in environments where global positioning system (GPS) signals are unavailable—such as indoors, underground, and underwater—has become a research hotspot in the Internet of Things (IoT). However, most existing traditional wireless positioning technologies can only achieve desirable results in environments with guaranteed line-of-sight (LoS) paths. In contrast, magnetoquasistatic (MQS) positioning technology has demonstrated strong competitiveness in the these environments due to its robust penetration capabilities. Based on recent state-of-the-art research, this article presents a comprehensive and exhaustive survey of MQS positioning, focusing on the foundation knowledge of MQS fields and the application of MQS positioning in various scenarios. Additionally, different existing and innovative solutions/methods for solving MQS positioning-related problems are presented, highlighting their pros and cons. Furthermore, this article categorizes MQS positioning according to their transmitter-receiver array combinations. Finally, critical challenges and future research prospects are presented. With this survey, we aim to provide a complete roadmap of previous and current research trends, identify research gaps, and suggest future research directions that will guide researchers in their subsequent advanced studies on MQS positioning.

Efficient trajectory planning in Autonomous Valet Parking (AVP) remains challenging due to multiple vehicle interactions and environmental complexities. Existing single-agent Reinforcement Learning (RL) approaches face challenges in balancing complexity, convergence, and knowledge efficiency, often resulting in increased collisions and longer travel times. To address these issues, this paper proposes a Bio-inspired Evolutionary Reinforcement Learning (BERL) framework for multi-vehicle collaborative trajectory planning, where each vehicle is modeled as a Fusion Architecture for Learning and Cognition Network (FALCON) agent based on Adaptive Resonance Theory (ART). The BERL framework comprises three core modules: 1)Meme Reinforcement Learning (MRL), which enables agents to learn independently and adapt to changing environments; 2)Expert-Guided Evolutionary Learning (EGEL), which facilitates knowledge transfer from expert agents to less experienced ones, enhancing coordination; and 3)Integrated Forgetting and Memory Optimization (IFMO), which optimizes memory use and reduces algorithm complexity. Additionally, the BERL framework supports model and sensor quality heterogeneity in the multi-vehicle trajectory planning scenario. Finally, we build an AVP Simulation (AVPS) platform to validate the performance of the proposed framework. Comprehensive simulation results demonstrate that the BERL framework improves success rate and parking efficiency by at least 15.7% and 16.7%, respectively, as compared to state-of-the-art algorithms. Additionally, the proposed IFMO module reduces the number of memes in the FALCON agent by 30.2% while maintaining stable performance.

Crowdsourcing is an effective method to alleviate the burden of conducting a site-survey procedure for localization tasks. However, crowdsourced data is typically inaccurately and scarcely annotated, rendering accurate localization a rather challenging problem. To alleviate this problem, we propose VRLoc, a deep reinforcement learning (DRL)-based unsupervised wireless localization framework using crowdsourced trajectory data. The proposed VRLoc primarily encompasses three components, i.e., a robust K-means (RKM) clustering method for generating a series of virtual reference points (VRPs), DRL for determining the physical layout for VRPs, and online localization based on VRPs. Specifically, the proposed RKM method employs a density-based approach for the initialization of cluster centers, rather than the commonly used random solution, yielding repeatable and reliable VRP generation results. To accurately determine the physical locations for VRPs, we develop a modified soft actor-critic (SAC)- based VRP layout method with multiple objectives, i.e., the connection topology among VRPs, the floor-plan information, and the near-field condition. Then, we effectively predict locations of target users by utilizing classification models to match the online collected samples with the VRPs annotated by physical locations. The proposed framework is advantageous in achieving high-accuracy unsupervised localization, with the VRPs bridging the unlabeled crowdsourced data and physical location space. Both experimental and simulation results demonstrate the effectiveness and superiority of the proposed VRLoc framework as an accurate and practical solution for unsupervised localization.

The industrial Internet of Things (IIoT) and network slicing (NS) paradigms are key enablers of the industrial revolution in current and future mobile networks. However, peer-to-peer (P2P) resource blocks (RBs) exchange to match supply and demand in sliced IIoT networks requires proper incentivization and renegotiations between the service providers (SPs). This paper models the business strategic interactions between seller and buyer SPs as a coalitional game in which sellers form coalitions to set RB prices and buyers join coalitions to determine their best-response RB demand. The aim is to maximize the profit of the seller coalition and minimize the expenses of the buyer coalition while jointly contributing to maximize system RB utilization. Due to the uncertainty of network traffic, we propose a coalitional game-guided multiagent reinforcement learning approach that takes the output of the coalitional game as the starting Nash equilibrium (NE) and computes the optimal price and demand strategies of the coalitions regardless of network condition changes. Simulation results and analysis prove the efficacy of the proposed approach in terms of optimizing seller and buyer coalition payoffs, as well as maximizing the overall RB utilization.

The heavy burden of fingerprint collection and annotation has become one of the biggest bottlenecks in wireless indoor positioning, particularly in the context of the Internet of Things (IoT). Fortunately, crowdsourcing can be leveraged to alleviate the fingerprint collection burden by harnessing the collective intelligence of crowdsourcing users. However, it is rather difficult to acquire an accurate positioning model based solely on training unlabeled crowdsourcing data. To overcome this problem, we propose a novel positioning model called ENvironment Aware Positioning (ENAP), utilizing unlabeled crowdsourcing trace data. The proposed ENAP mainly consists of three steps, i.e., transforming the unlabeled crowdsourcing trace data into a cluster space, mapping the cluster space into the positioning space, and continuously updates the positioning model in an unsupervised manner. To enhance the performance and robustness against device heterogeneity of crowdsourcing users, we propose a novel clustering scheme for space transformation by adaptively fusing multiple signal features. Then, to ensure long-term positioning stability and continual environmental aware capability, we incorporate a dynamic replay memory into ENAP that enables the unsupervised online updating of positioning models, distinguishing our proposal from most existing positioning models. Simulation and experimental results demonstrate the effectiveness and superiority of the proposed ENAP approach as a practical and efficient solution for wireless indoor positioning in the IoT era.

Source localization within wireless sensor networks (WSNs) is one of the critical technologies in the Internet of Things (IoT). As the number of network nodes increases, so does the amount of data and computational requirement. It is imperative to introduce edge computing. However, there are still two issues when running existing wireless location algorithms on edge nodes: 1) conventional low-complexity approaches are easily affected by the bias generated in complex environments, leading to low locating accuracy and 2) the optimization algorithms considering the bias have good performances, but they are calculation-efficiency low on edge nodes. This study proposes a computationally efficient and high-precision location method to tackle the troubles. Precisely, we first introduce our previous research to construct a bias-considered nonconvex problem with a linear objective. Then, we propose an angle-assisted Taylor series with zero truncation error to linearize the second-order cone (SOC) constraint in the established problem. Next, we resort to the mini-max criterion to eliminate the angular uncertainty and get a robust linear programming (LP) problem with an optimal solution. So far, we have obtained a convex problem of low complexity. To ensure the calculated efficiency of the proposed problem on edge nodes, we proceed to give the solving process of the problem. Moreover, we provide a constraints tracking mechanism to reduce the number of iterations in the solution procedure, improving computational efficiency. Simulations and experiments demonstrate that the proposed method with similar locating accuracy to state-of-the-art optimization algorithms exhibits much higher computing efficiency on edge nodes.

While most existing transfer learning methods map the source domain and the target domain data into a common space by sharing the model, the specific knowledge with private properties may be lost. To address this problem, we propose a transfer learning model with shared and specific structures for synthetic aperture radar (SAR) target recognition in this paper. Firstly, we design a convolutional neural network model with the shared subnetwork and the specific subnetworks. The shared subnetwork transforms the common information of the source and the target domain data into a common subspace, while the specific subnetworks transform the private information into the specific subspaces separately. The proposed model does not only extract the similar structure of the different domains but also reserve their private properties. Secondly, by considering the different contributions of the common-subspace features and the specific-subspace features for the final classification, we design adaptive feature weights to compute the domain features. Lastly, in addition to the classification loss, the model is trained by reducing distribution discrepancy loss, which establishes a knowledge transfer bridge from the labeled source domain to the unlabeled target domain. The experimental results verify the effectiveness of the proposed method.

Traditional fingerprint-based positioning approaches work well on static data; they cannot handle scenarios where the data distribution, the feature space and even the signal source evolve over time, which are ubiquitous in real-world applications. One straightforward approach for circumventing these difficulties is to repeat labeled data calibration for maintaining an up-to-date fingerprint database, which is usually infeasible or expensive in large-scale indoor environments. In this paper, we propose a Long Short-Term indoor Positioning (LSTP) framework that enables adaptation at different time scales with low human-effort, and thus extends the effectiveness of existing fingerprinting techniques for a more generalized environment. Specifically, LSTP mainly considers the distribution discrepancy caused by continuous environmental dynamics in short-term positioning and the feature space heterogeneity in long-term positioning. To address the first challenge, we design an incremental ensemble localization model which leverages multiple source classifiers to resolve distribution differences in an online manner. To address the second challenge, we seek to borrow knowledge learned from an earlier time period with plenty of labeled samples for the current time period, thus reducing the required number of new calibration samples. By fully capturing the transferable spatial information across different time periods with multi-level constraints (sample, feature, and model levels), we can study a discriminative domain-invariant space from which we can make better predictions. The experiments on three real-world datasets demonstrate the superiority of the proposed framework, which outperforms the state-of-the-art systems by 18% in mean accuracy.

To probe an accurate diagnosing approach for synchronous generator (SG) with rotor winding inter-turn short-circuit, a novel online monitoring and detecting method relying on the [Formula: see text]-support vector regression ([Formula: see text]-SVR) machine was proposed, and its effectiveness was further verified by the micro-synchronous generator dynamic simulation. Terminal voltage, active and reactive power of SG were selected as input variables for a novel prediction model based on the [Formula: see text]-SVR, and field current was selected as an output variable of the prediction model. The structures and parameters of the field current prediction model were optimized with the particle swarm optimization (PSO) algorithm and training samples, then the prediction model was established and the field current prediction got under way. By comparing the predicted field current with the corresponding online measured field current, inter-turn short-circuit of rotor winding in SG could be detected sensitively once its absolute value of the prediction relative error exceeded a specific threshold. The micro-synchronous generator dynamic simulation indicated that the proposed online detecting approach based on the [Formula: see text]-SVR machine overcame the shortage of the back-propagation (BP) diagnosis method for misdiagnosis, and its accuracy, sensitivity and threshold setting range of the diagnosis method was the most prominent among these diagnosis methods such as the BP diagnosis method, the Bayesian regularization back-propagation (BRBP) diagnosis method and the [Formula: see text]-support vector regression ([Formula: see text]-SVR) diagnosis method.

This paper presents a novel model for ultrasonic defect identification relying on multi-feature fusion and multi-criteria feature evaluation (MFF-MCFE). Based on feature extraction, feature selection, pattern recognition and data fusion algorithm, this model analyzes ultrasonic echo signal data from single-probe ultrasonic inspection, and based on wavelet packet transform (WPT), empirical mode decomposition (EMD) and discrete wavelet transform (DWT), the main features from the collected ultrasonic echo signals are also extracted. These features are also evaluated by means of Representation Entropy (RE), Fisher’s ratio (FR) and Mahalanobis distance (MD), and the results are fused with Dempster–Shafer (D-S) evidence theory and the corresponding feature subsets are formed according to the fusion result. The support vector machine (SVM) is used as the classifier to recognize the defect signal, and the subsequent classification results are integrated by D-S evidence theory, which leads to the final recognition results. On this basis, a series of experiments were carried out to compare the performance of the developed model with that of the models using single feature sets and single feature evaluation criterion. Meanwhile, the principal component analysis (PCA) was also involved in the corresponding comparative analysis. The experimental results showed that this model is suitable for the identification and diagnosis of pipeline defects, and its classification accuracy could be reached up to 96.29% with stronger robustness and stability.

Transfer learning algorithms (TLAs) are often used to solve the distribution discrepancy issue in fingerprint-based indoor localization. However, existing TLAs cannot react well to real time changes in the environmental dynamics of the target space due to three remarkable shortcomings: a) redundant knowledge in source domain may lead to “negative transfer”; b) the required target domain samples to calculate the distributions are unrealistically feasible for real-time positioning; c) they cannot transfer knowledge efficiently across domains with heterogeneous feature spaces. In this paper, we propose TransLoc, a heterogeneous knowledge transfer framework for fingerprint-based indoor localization, which can perform knowledge transfer efficiently even with only one sample in the target domain. Specifically, we first refine the source domain according to the target domain by removing redundant knowledge in the source domain. Then, we derive a cross-domain mapping, which transfers the specific knowledge of one domain to another domain, to construct a homogeneous feature space. In this new feature space, the transfer weights are computed for training a classifier for target location prediction. To further train the framework efficiently, we combine the mapping and weights learning into a joint objective function and solve it by a three-step iterative optimization algorithm. Extensive simulation and real-world experimental results verify that TransLoc not only significantly outperforms state-of-the-art methods but is also very robust to changing environment.

Images from different aspect views for one target, known as multiple views, are widely applied to improve synthetic aperture radar (SAR) target recognition. However, most of existing multi-view methods have strict constraint on the angle interval among multiple views. In this paper, a new multiview fusion method free from interval limitation using expectation maximization (EM) is explored for SAR image classification. Firstly, we apply convolutional neural network (CNN) to extract features effectively owning to its powerful ability of feature learning and then obtain the classification probability. Secondly, Multi-view Label Set (MLS) is automatically constructed from multiple views according to the probability and finally we use EM algorithm to classify SAR images intelligently. It is worth noting that the proposed method can be used flexibly according to the number of perspectives obtained and without angle interval constraint among multiple views. Experiments demonstrate that the proposed method has better recognition performance than some state-of-the-art methods on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset.

It is notable that localization accuracy using received signal strength (RSS) fingerprints solely is very vulnerable to dynamic environments. Utilizing multiple fingerprints gleaned from RSS for localization is a propitious strategy to overcome the RSS susceptibility. Brimful utilization via fusing multiple fingerprint functions which supplement each other are not harnessed by existing fusion-based techniques, resulting in low localization accuracy. This paper presents a novel and robust WiFi localization modus operandi by fusing DerIvative Fingerprints of RSS with MultIple Classifiers (DIFMIC). DIFMIC first constructs a multiple fingerprints group by gleaning hyperbolic location fingerprint (HLF) and signal strength differences fingerprint (DIFF) from RSS fingerprints. Then, it obtains Multiple Fingerprints Trained Classifiers (MFTCs) via training each basic classifier with each fingerprint. To fully leverage the inherent supplementation among fingerprints and classifiers, a two-layer fusion profile (weights) joint optimization algorithm with multiple constraints is proposed. We also propose a Fusion Profile Selection (FPS) algorithm to intelligently choose fusion weights from the two-layer fusion profile for a more accurate localization. DIFMIC shows more leverage in combining multiple information, thus exhibiting better robustness in WiFi positioning. Results from our experiments reflect that DIFMIC performs better than other existing methods in real environments.

With the growth of WLAN infrastructure, received signal strength (RSS) fingerprint-based WiFi indoor positioning systems have received considerable attention recently. Some existing RSS fingerprint-based localization methods estimate locations by directly matching an online testing sample with an offline database, and thus show low accuracy because RSS is known to be vulnerable to variations caused by changing environment and heterogeneous hardware. To overcome the above drawbacks, we propose an expectation maximization indoor localization approach by leveraging supporting set (EMSS). In the offline phase, we first divide a positioning area into ${G}$ grid points and index each grid by a label. All the indices of grid points form a label set $ {\Psi =\{1,2,\ldots, G\}}$ . Then, we collect the RSS fingerprints to construct an offline database for all labeled grid points. In the online phase, given an online RSS testing sample, we first construct a supporting set (SS), which is a subset of $ {\Psi }$ , selected by the similarity between the online RSS sample and offline database. So, SS is a latent space that likely includes the true label (location) of the user. Based on the SS, we then derive an expectation maximization (EM) algorithm by incorporating the fingerprint quality into the estimation of the true label. EM can intelligently estimate the location of the user by evaluating the fingerprint quality of SS. Furthermore, we propose an optimal size selection algorithm based on Bayesian information criterion to adaptively determine the size of SS. Our method can effectively mitigate the impacts of changing environment and heterogeneous hardware without fingerprint and hardware calibrations, and can thus be practically applied. Experimental results verify that EMSS performs significantly better than some existing fingerprint-based methods.

Fusing the predictions of multiple received signal strength (RSS)-based classifiers is an efficient strategy to mitigate the impact of the fluctuation of the RSS. However, most of the existing fusion methods exhibit two remarkable shortcomings: 1) they need to train and store offline weights by the supervised learning and 2) they directly fuse the so-called candidate location set (CLS), which is collected from the most likely location estimate of each classifier (location with the largest probability of being the true location predicted by the classifier), and thus do not fully leverage the knowledge of classifiers. In general, the fluctuation of RSS does not guarantee the location predicted by each classifier with the highest probability to be the true location, thus leading to severe performance degeneration of the existing fusion methods. To overcome the above shortcomings, we propose an accurate WiFi localization framework by unsupervised fusion of an extended CLS (ECLS). First, we train multiple classifiers by only using RSS fingerprints in the offline phase. In the online phase, instead of collecting the CLS from the trained classifiers, we construct an ECLS by augmenting CLS with other location estimates (locations with predication probability greater than a certain threshold) from each classifier. As compared with the CLS, ECLS provides a bigger fusion space that likely includes the true location of the user. Furthermore, an unsupervised fusion localization algorithm based on the ECLS is derived from the joint optimization of weights and the location of the user. Furthermore, a point of inflection searching algorithm is also proposed to intelligently construct the ECLS. Real experimental results show that our proposed algorithm is more robust to changing environments and model errors, and can significantly improve the localization accuracy without any fingerprint and hardware calibrations.

The main drawback for large-scale applications of WiFi-based localization is the varying characteristics of received signal strength (RSS), which degenerates the localization performance seriously. To mitigate the variation problem, we propose a hybrid fingerprint quality evaluation model (HFQuM) for accurate WiFi localization. HFQuM can intelligently determine the location of a user by evaluating the hybrid fingerprint quality in different subareas, that is a high fingerprint quality indicates that the frequently occurred location label is more likely to be true. To achieve this, in the offline phase, instead of only collecting RSS fingerprints, we construct a WiFi-based group of fingerprints (GOOFs) consisting of RSS, signal strength difference (SSD), and hyperbolic location fingerprint (HLF). Given an RSS testing sample of a user at an unknown location in the online phase, we first construct the multiple supporting sets (MSSs), including a sample space and a label space, selected by the similarity between the online sample and the GOOF. Based on the MSS, HFQuM is able to estimate the user's location as well as subareas and their hybrid fingerprint quality simultaneously by jointly modeling the process of generating the sample space and label space. To further reduce the computational complexity, HFQuM employs an access point (AP) selection algorithm to exclude redundancy APs. Experimental results in a typical library environment verify the superiority of HFQuM in terms of localization accuracy as compared with other existing fingerprint-based methods.

Indoor localization is becoming critical to empower Internet of Things for various applications, such as asset tracking, autonomous parking, virtual reality, context awareness, condition monitoring, geolocation, smart manufacturing, as well as smart cities. It is well known that indoor localization based on some single fingerprints is rather susceptible to the changing environment. The efficiency of building single fingerprints from one localization system is also low. Recently, we first proposed a group of fingerprints (GOOF) based localization to improve the efficiency of building fingerprints, and then proposed an efficient fusion algorithm, namely, multiple classifiers multiple samples (MUCUS), to improve the accuracy of localization. However, the main drawbacks of MUCUS are the low localization efficiency and low accuracy when all classifiers show poor performance simultaneously. In this paper, based on the aforementioned GOOF, we propose a sliding window aided mode-based (SWIM) fusion algorithm to balance the localization accuracy and efficiency. SWIM first adopts windowing and sliding techniques to improve the localization efficiency, and then obtains a more accurate estimate by minimizing the entropy of multiple classifiers or multiple samples. This can guarantee our estimator to be robust to changing environment and larger noise level. We demonstrate the performance of our algorithms through simulations and real experimental data via two universal software radio peripheral platforms.

Indoor localization is becoming critical to empower Internet of Things for various applications, such as asset tracking, geolocation, and smart cities. Wi-Fi-based indoor localization using received signal strength (RSS) has drawn much attention over the past decade because it does not require extra infrastructure and specialized hardware. It is well known that the localization accuracy using RSS is rather susceptible to the changing environment. Localization by fusing multiple fingerprint functions of RSS is a promising strategy to overcome the above drawback. However, the existing fusion techniques cannot make full use of the intrinsic complementarity among multiple fingerprint functions. It also fails to exploit the knowledge obtained in the offline phase and thus shows low accuracy in the complex environment. This paper proposes a knowledge aided adaptive localization (KAAL) approach by using a global fusion profile (GFP) to mitigate the above shortcomings. First, we propose a GFP construction algorithm by minimizing position errors over all fingerprint functions with weight constraints in the offline phase. Based on the knowledge from GFP and the trained multiple fingerprint models, we then derive two KAAL algorithms, namely, multiple function averaging and optimal function selection, to achieve highly accurate localization results. Experimental results demonstrate that our proposed localization approach is superior to the existing methods both in simulated and real environments.

An improved direction-of-arrivals (DOAs) estimation via phase information of sparse solution is presented in this paper. Unlike the conventional sparse source localization approach using the amplitude of sparse solutions only, through a special partition of the receiving data of the sensors, the phase information of the available sparse solutions is also extracted to estimate DOAs. For the true DOAs exactly on the grids which are used to generate the over-complete dictionary, the performance of our method is close to the conventional sparse source localization method. For the true DOAs that are not on the grids, our method is far superior to the conventional method, as demonstrated by several simulation results.