Ultra-Wideband (UWB) technology enables high-precision positioning in underground subway environments with poor GNSS signal coverage, but ranging measurements often suffer from data loss due to occlusions, multipath effects, and ranging failures. While recent studies have focused on imputing missing measurements using temporal modeling, they often overlook spatial factors such as anchor group vary and train motion patterns. To address this limitation, this paper proposes a novel approach for imputing missing UWB ranging data by jointly exploiting temporal dynamics and spatial constraints. A mask matrix is introduced to model data availability, and a spatio-temporal optimization framework is established to guide the imputation process. Based on this framework, a Dynamic Spatio-Temporal Missing Data Imputation Network (DSTMIN) is developed for robust localization in complex subway environments. In particular, DSTMIN leverages missing-aware attention and adaptive graph fusion, significantly enhancing its robustness to large missing blocks and non-random data loss. Simulation results demonstrate that DSTMIN outperforms state-of-the-art models, such as STGCN and GRIN, under various missing rates. Under a 40% ranging data missing ratio, DSTMIN reduces the root mean square error (RMSE) by 18.6% and 16.3% compared with STGCN and GRIN, respectively.

Intelligent radio maps (IRMs) have emerged as a critical enabler for next-generation wireless networks, offering comprehensive spatiotemporal awareness of the electromagnetic environment with limited sensing resources and low computational overhead. They play a crucial role in enhancing spectrum efficiency, enabling intelligent resource allocation, supporting anti-jamming communications, improving interference management, and facilitating environment-aware networking. This paper presents a systematic overview of how to construct high-quality IRMs. We first introduce six evaluation metrics aligned with practical deployment requirements and evolving wireless network demands. Guided by these metrics and recent advances in artificial intelligence (AI), we provide an in-depth review of spectrum sensing approaches and state-of-the-art methods for spectrum inference. We further explore the intrinsic connections between these two steps and propose an integrated sensing–inference construction scheme. Extensive experiments demonstrate that the integrated scheme achieves superior IRM construction performance under sparse sensing, validating its practical potential for future wireless networks.

With the rapid growth of Internet of Things and proliferation of handheld smart devices, mobile crowdsensing has been regarded as an effective sensing paradigm due to its high scalability, low cost, and wide coverage. In this paper, we study hybrid task assignment where semi-opportunistic and participatory users co-exist for task executions while tasks are delay sensitive and have heterogeneous qualities. The design objective is to maximize the total quality of completed tasks subject to a total budget shared by both types of users. We formulate this problem as an integer programming problem. We propose an efficient hybrid users based task assignment algorithm (referred to as HU-TSA), which works in an iterative way as follows. It first selects the top n (initially, n = 1) semi-opportunistic users in terms of quality-cost ratio for task assignment. It then clusters the remaining tasks into different regions based on their closeness and then performs utility based optimized user-region binding and standardized task density based path planning for the participatory users. It repeats the above process over all possible values of n to seek an optimal budget splitting between the two types of users for improved performance. We present the detailed design description of HU-TSA and deduce its computational complexity. Extensive simulations are carried out and the results show the effectiveness of HU-TSA by comparing with existing algorithms.

Fingerprint-based indoor localization is one of the most promising solutions for various Intelligent Internet of Things (IIoT) systems. However, recent studies show that the key design challenges of current fingerprint-based localization techniques come from the following three aspects: 1) temporal variation caused by various patterns of IIoT device operations and stochastic fluctuation of wireless signals, 2) spatial unevenness of collected RSSI samples due to complex multi-floor environments, and 3) high feature sparsity of collected RSSI samples in large areas. To address these challenges, we present a localization architecture for multi-floor indoor localization in multi-building environment and accordingly propose a fingerprint-based localization method (referred to as GrowNetLoc) based on Gradient Boosting Neural Network (GrowNet) and Long Short-Term Memory (LSTM) network. Regarding building/floor identification, the gradient ensemble model GrowNet is utilized for extracting the mapping relationship between uneven RSSI samples and building/floor indices. Regarding location estimation, LSTM network is adopted as one layer of base learner to extract temporal features of RSSI samples, and a gradient boosting strategy is further used for overcoming the sample sparsity issue and improving the location estimation performance. Extensive experiments are conducted on real datasets and the results demonstrate that GrowNetLoc has superior localization accuracy and robustness performance compared with the existing methods.

Cloud-aidedmobileedgenetworks (CAMENs) allow edge servers (ESs) to purchase resources from remote cloud servers (CSs), while overcoming resource shortage when handling computation-intensive tasks of mobile users (MUs). Conventional trading mechanisms (e.g., onsite trading) confront many challenges, including decision-making overhead (e.g., latency) and potential trading failures. This paper investigates a series of cross-layer matching mechanisms to achieve stable and cost-effective resource provisioning across different layers (i.e., MUs, ESs, CSs), seamlessly integrated into a novel hybrid paradigm that incorporates futures and spot trading. In futures trading, we explore anoverbooking-drivenaforehandcross-layermatching (OA-CLM) mechanism, facilitating two future contract types: contract between MUs and ESs, and contract between ESs and CSs, while assessing potential risks under historical statistical analysis. In spot trading, we design two backup plans respond to current network/market conditions: determination on contractual MUs that should switch to local processing from edge/cloud services; and anonsitecross-layermatching (OS-CLM) mechanism that engages participants in real-time practical transactions. We next show that our matching mechanisms theoretically satisfy stability, individual rationality, competitive equilibrium, and weak Pareto optimality. Comprehensive simulations in real-world and numerical network settings confirm the corresponding efficacy, while revealing remarkable improvements in time/energy efficiency and social welfare.

Wide dissemination of unverified claims has negative influence on social lives. Rumors are easy to emerge and spread in the crowds especially in Online Social Network (OSN), due to its openness and extensive amount of users. Therefore, rumor detection in OSN is a very challenging and urgent issue. In this paper, we propose an Attention-Residual network combined with CNN (ARC), which is based on the content features for rumor detection. First, we build a data encoding model based on word-level data for contextual feature representation. Second, we propose a residual framework based on fine-tuned attention mechanism to capture long-range dependency. Third, we apply convolution neural network with varying window size to select important components and local features. Experiments on two twitter datasets demonstrate that the proposed model has better performance than other content-based methods both in rumor detection and early rumor verification. To the best of our knowledge, we are the first work that utilize attention model in conjunction with residual network on rumor detection.

Mobile crowdsensing is a novel paradigm by exploiting mobility, sensing, computation, and communication capability of smart devices. In this paper, we study data offloading problem for mobile crowdsensing in opportunistic social networks. In this scenario, mobile users can upload sensing data directly via cellular networks using various data plans. A mobile user can also resort to another user for data offloading by forwarding sensing data to that user using short-range communications (when they encounter). To minimize total data uploading cost while meeting given uploading deadlines, data plan assignment for users and data forwarding strategy when two users encounter should be elaborately designed. In this paper, we use Benders decomposition algorithm to solve offline data plan assignment problem. Then we propose two algorithms including progress- balanced algorithm and social-aware forwarding algorithm to solve online data forwarding problem. Simulation results show that data offloading between users can largely reduce the total data uploading cost. Simulation results also show that the performance of our proposed online algorithms is close to the offline optimal solution.

With the widespread of smart devices, mobile crowdsensing has become an attractive way to perceive and collect sensing data. In this paper, we focus on studying AP-assisted task assignment in mobile crowdsensing. The objective is to effectively reduce the average or worst-case makespan of tasks. We focus on a scenario that a task requester needs the assistance of mobile users for task accomplishment while they can meet directly or via APs in an opportunistic manner. We model the crowdsensing system and then formulate the problems under study. We then propose an AP-assisted average makespan sensitive online task assignment (AP-AOTA) algorithm and an AP-assisted largest makespan sensitive online task assignment (AP-LOTA) algorithm. In the proposed algorithms, task assignment at each step considers both the inter-encountering time between requester and each user and that between them while going through APs. We present design details of the proposed algorithms. We derive their computational complexities to be O(mn2), where m is the number of tasks and n is the number of users. Finally, trace-driven simulation results show that the proposed algorithms outperform existing work.

Mobile crowdsensing is a new sensing paradigm exploiting potential of crowds to collect data, which has various advantages over traditional sensor networks such as low cost, high coverage, and high mobility. Privacy preservation is a crucial issue in mobile crowdsensing because worker privacy might be exposed if workers share their location information to service platform or other workers. In this paper, we assume workers can determine their own privacy preservation levels and they do not need to upload their location information to the platform or share to other workers for sensing behavior coordination. Moreover, workers move to task locations to collect sensing data in a distributed manner. We accordingly propose a privacy-aware online task assignment framework to achieve high task coverage. In this framework, spatial task-application information in previous cycles is used to estimate worker density and an incentive pricing mechanism is designed to guide workers to collect sensing data in low-worker-density areas. We present detailed mechanism design. Extensive simulation results show that our proposed solution has much better performance than the baseline mechanism.

Mobile crowdsensing is a sensing paradigm such that mobile users need to move to task locations to perform sensing tasks. In this paper, we focus on studying the task assignment problem of eco-friendly mobile crowdsensing which aims to minimize carbon emissions while meeting various resource limits including task deadlines and transportation constraints. We first describe the eco-friendly mobile crowdsensing system model and formulate the task assignment problem. Then we divide the problem into two subproblems including selection of best transportation type and user-task matching. We model the user-task matching as unbalanced minimum-cost bipartite matching, transform the problem into balanced maximum-weight bipartite matching, and use Kuhn-Munkres algorithm to obtain the optimal solution. Extensive simulations are conducted and the results show the efficiency and effectiveness of our proposed solution.

Smart devices with a rich set of low-cost sensors enable a new sensing paradigm called mobile crowdsensing. In mobile crowdsensing, tasks are distributed at a variety of locations. Mobile users travel through different task locations to perform different tasks. The diversity of task locations and user trajectories makes the optimal scheduling problem intractable. In this paper, we mathematically formulate the optimal task scheduling problem as a continuous path planning problem, which is known to be NP-hard. Then we propose two online heuristic algorithms to maximize the task quality improvement for each newly arriving user. These algorithms work in a hop by hop manner for task selection and adopt different measures and strategies including: (1) ratio of task quality increment and travel cost and (2) task spatial density. We present detailed algorithm design and deduce their computational complexity. Extensive simulation results show that our algorithms outperform existing work.

In this paper, we study how to effectively reduce the average end-to-end (E2E) packet delay in backpressure based scheduling in wireless multi-hop networks. We accordingly propose a virtual gradient based back-pressure scheduling algorithm, referred to as VBR. In VBR, intentional virtual queue, whose length (called virtual gradient) depends on the distance to destination, is first built at nodes in a network in the network configuration phase. In this way, virtual gradient is established at nodes in the network. In the network operation phase, the scheduling decision at each node needs to jointly consider both real queue length and virtual queue length. Simulation results show that VBR can obtain significant performance improvement on back-pressure based routing and scheduling, in terms of packet delivery ratio and average E2E delay.