Advanced Persistent Threats (APTs) have become increasingly sophisticated and covert, necessitating the acquisition of an overall view of the rapidly evolving cyber threat landscape by security defenders. However, integrating threat intelligence from diverse sources poses significant challenges due to limited labeled data and noise interference. To address the requirement for the early detection of APT attacks, this paper introduces a lightweight framework named Warning-Graph, based on threat intelligence modeling. Warning-Graph leverages a limited set of IoCs to infer the type of ongoing APT attack. Initially, attack-related infrastructure nodes are modeled as a heterogeneous information network. Subsequently, heterogeneous graph contrastive learning is employed for pre-training. Two asymmetric graph encoders are constructed to obtain node embeddings without the need to generate negative samples or labeled data. In addition, a loss function based on the information bottleneck is specifically employed to reduce the noise in the original graph. In downstream tasks, multiclass classifiers are trained using embedding representations with fewer labeled samples. Experimental results demonstrate that the proposed framework achieves a 3- to 5-point increase in identification performance for APT attack types compared to baselines, while utilizing fewer labeled samples.

Vehicle accidents result in a large number of fatalities and substantial economic losses annually. A significant portion of vehicle accidents are attributed to driver negligence, particularly drowsy driving (often caused by fatigue), which can seriously weaken drivers' alertness in steering control, lane keeping, and maintaining a safe distance from other vehicles. Thus, effective drowsiness detection is essential for preventing accidents. The fatigue state of drivers can be inferred from videos captured by in-vehicle cameras. To achieve accurate and rapid detection, both personal and common fatigue features should be learned from historical driving videos and jointly exploited for state recognition. In this paper, we propose an edge AI-enabled drowsiness detection framework. We design a dual-branch architecture for feature fusion that integrates the local models to capture personalized fatigue features and the regional models to extract common fatigue features, thereby improving the detection accuracy while reducing detection latency. Furthermore, to lower computational overhead, we identify key frames in videos and represent facial landmarks with relative polar coordinates, which accelerates feature extraction. Extensive experiments demonstrate that our approach performs well in terms of accuracy, AUC, and F1 score (0.9986, 0.9999, and 0.9986, respectively). Moreover, only 85 frames (45 I-frames and 40 P-frames) are required per detection instead of the original 795 frames, resulting in a detection delay of approximately 517 ms.

Unmanned Aerial Vehicles (UAVs) have recently been widely used in various fields. However, both cooperative charging scheduling and insufficient charging facility problems in UAV charging scenarios have been rarely studied. This paper studies parallel cooperative charging scheduling of UAVs. We adopt cooperative charging to reduce the total cost and parallel scheduling to enable UAVs can be charged even if the number of UAVs is more than the number of charging facilities. We formulate the Parallel Cooperative Charging Scheduling for UAVs Problem (PCCSUP) for optimizing the total cost of whole charging system. We first investigate the special case of PCCSUP with single charging station, and use the approximation algorithm for Uniform Parallel Machines Scheduling Problem (UPMSP) to solve the special case. Then, a greedy approach based approximation algorithm is proposed to solve the PCCSUP, where we use the approximation algorithm for UPMSP to obtain the charging arrangements and the Set Covering Problem (SCP) optimization framework to obtain the charging groups. The results of extensive simulations demonstrate that our algorithm can reduce up to 59.81% total cost compared with the benchmark algorithms. Finally, we discuss and design the algorithms for three related problems: PCCSUP with different arrival times, PCCSUP withK-anonymity, and charging arrangements for excluded UAVs.

Recently, secure and reliable identity recognition in mobile phones has become critical due to the rising cyber threats. Traditional identity recognition methods, such as passwords, are prone to brute-force attacks and phishing, while the existing biometric recognition, such as face recognition, voice recognition, and fingerprint recognition, could be confronted with the issues of leakage and variability of the identity features, making them easy to replicate or degrade over time due to varying ages and environmental factors. To this end, we propose the wrist-rolling motion recognition (WRMR) method. WRMR leverages the unique motion patterns that require the sensors embedded in mobile phones. These motion patterns rely on the complex muscle memory and coordination, which makes the motion patterns highly individual and difficult to imitate, reducing the risk of the identity features leakage. Moreover, the wrist-rolling motion patterns are less affected by the ages and emotional states of users, ensuring the long-term consistency and reliability. In WRMR, we specially introduce a Transformer-based model, termed patch time series transformer with the temporal-spatio attention (PatchTST-TSA), which accurately classifies the time series data generated by the wrist-rolling motion. PatchTST-TSA enhances the original Transformer model structure by incorporating the patch embedding and temporal-spatio attention (TSA), thus effectively capturing the local dependencies and extracting the temporal-spatio features. Extensive experiments demonstrate that PatchTST-TSA significantly improves the classification performance and noise resilience. WRMR presents a robust and accurate solution for the mobile phone unlocking, and highlights the potential of the time series classification in the identity recognition.

Depression has currently become a serious social problem worldwide. However, the need for experienced doctors and tedious medical examinations greatly increases the inconvenience in diagnosing the depression. A convenient depression diagnosis method can significantly improve the medical experience of depression patients, and can greatly reduce the workload of doctors. In this paper, a Convenient Depression Diagnosis method based on Audio-Visual Deep Learning (CDD-AVDL) is proposed. CDD-AVDL exploits the videos of testers reading a specially-designed text, and note that the videos contain many subconscious human reactions (e.g., micro expressions, voice variations), which are difficultly affected by the artificial interventions, thus enabling the depression diagnosis results more accurate. In CDD-AVDL, the source features are first extracted from audios and visuals, and then the time-sequential features are extracted. Finally, a full connection layer and a convolution layer fusion are responsible for fusing the audio-visual features to yield the depression probabilities. Extensive experiments and clinical tests show that CDD-AVDL outperforms the state-of-the-arts in terms of the accuracy of depression diagnosis. Moreover, the data collection manner in CDD-AVDL is convenient, and the training cost of CDD-AVDL is very low.

With the rapid development of Internet of Things and the increasingly popularized smart devices, the biometric recognition becomes a crucial component of facilitating some basic activities of daily living, and the biometric recognition has the outstanding advantages in terms of reliability and convenience. Various biometric characteristics have been applied to realize the biometric recognition for basic human activities, and new biometric characteristics are worth exploring to further enhance the convenience of our daily lives. This paper explores a new biometric characteristic (the wrist-rolling motion). By taking the mobile phone unlocking as the typical application of the biometric recognition, we verify that the wrist-rolling motion can become an available biometric characteristic with the aid of our designed deep learning model termed RTimesNet. Specifically, RTimesNet is composed of TimesBlocks, decomposition modules, and a multi-head ProbSparse self-attention module, and it exploits the periodicity of wrist-rolling motion to extract the time series features. TimesBlocks extract the features hidden in the wrist-rolling motion, and the decomposition modules decompose the output data of TimesBlocks into the trend-cyclical data and seasonal data, which are then evenly divided and inputted into the multi-head ProbSparse self-attention module for concatenation. In addition, a federated learning manner is adopted for the motion-based biometric recognition, thus avoiding the exchange of local data and protecting the privacy of users. Extensive experiments have been conducted, and the results demonstrate that the wrist-rolling motion can become an available biometric characteristic. Compared with other biometric recognition methods, our proposed method shows a faster unlocking speed and requires less data storage with a satisfactory biometric recognition accuracy.

The Domain Name System (DNS) is a critical Internet service that translates domain names into IPs, but it is often targeted by attackers, posing a serious security risk. Graph-based models for detecting malicious domains have shown high performance but are vulnerable to adversarial attacks. To address this issue, we propose RMD-Graph, which is characterized by its ability to resist adversarial attacks and its low dependency on labeled data. A dual denoising module is specifically designed based on two autoencoders to generate the reconstructed graph, where SVD, TOP-k and reconstruction loss are introduced to enhance the denoising capability of autoencoders. Subsequently, residual connections are employed to generate an optimized graph that retains essential information from the original graph. The reconstructed graph and the optimized graph are then utilized as two views for graph contrastive learning, thereby achieving an self-supervised representation learning task without labels. In the downstream malicious domain detection, the denoised node representations are employed for machine learning classification. Extensive experiments are conducted on publicly available DNS datasets, and the results demonstrate that RMD-Graph significantly outperforms known baseline methods, especially in adversarial scenarios.

The deployment of Unmanned Aerial Vehicle (UAV) cluster is an available solution for object detection missions. In the harsh environment, UAV cluster could suffer from some significant threats (e.g., forest fire hazards, electromagnetic interference, and ground-to-air attacks), which could lead to the destruction of UAVs and loss of data. To this end, we propose a Hierarchical Federated Learning Framework for Object Detection (HFL-OD) to enhance the robustness of UAV cluster conducting object detection missions. In HFL-OD, UAVs are grouped through a Three-Dimensional (3D) graph coloring method, and an intragroup backup mechanism is provided to prevent the data loss caused by the destruction of UAVs. Besides, a dynamic server selection mechanism deals with the potential destruction of servers (cluster server and group servers) by adaptively reassigning the server roles. To further improve the robustness and mission efficiency of UAV cluster, a twotier federated learning framework is introduced to make a proper trade-off between object detection accuracy and communication/computational overhead. This framework is built on the concept of hierarchical federated learning by implementing both intragroup parameter aggregation and global parameter aggregation. Extensive simulations and comparisons demonstrate the superior performance of our proposed HFL-OD, i.e., the robustness of UAV cluster conducting object detection missions can be significantly improved, and the communication/computational overhead is effectively reduced.

At present, Unmanned Aerial Vehicle (UAV) swarm has been extensively applied in various fields. In the application of detection and localization of electronic signals, some UAVs could become disabled due to some abnormal events (e.g. electromagnetic interference and battery electricity exhaustion), and the topology connectivity of UAV swarm could be impaired, i.e., the topology of UAV swarm could be partitioned. For the topology recovery issue, we first propose Robust Topology Recovery Algorithm of UAV swarm (RTRA) to recover the topology connectivity of UAV swarm and enhance the topology robustness (reduce the number of potential topology recoveries in future) by relocating some UAVs to new positions with shortest flight distance. Furthermore, we note that the relocated UAVs are easy to exhaust the battery electricity and fail due to the extra flight movements for the topology recoveries, which affects the topology robustness. To this end, we present Cascading Robust Recovery Topology Algorithm of UAV swarm (CRTRA), which adopts a cascading movement strategy to share the flight movements among multiply relocated UAVs, thus avoiding the battery electricity exhaustion of the relocated UAVs. Extensive simulations and comparisons demonstrate that our proposed CRTRA can effectively recover the topology connectivity of UAV swarm while enhancing the topology robustness and shortening the flight distance of relocated UAVs, and CRTRA is especially suitable for some missions such as the detection and localization of electronic signals where UAVs are prone to fail.

Link prediction refers to the prediction of the potential relationships between nodes through exploring the evolution of the historical network topologies. In mobile social networks, the topologies change frequently due to the appearance/disappearance of nodes over time, and the links between nodes are typically very sparse (i.e., mobile social networks are with strong sparsity), which could affect the accuracy of link prediction in mobile social networks seriously. Therefore, this paper proposes a deep learning based Link Prediction Method against Strong Sparsity (LPMSS). LPMSS integrates the graph convolutional network output with encounter matrices to mitigate the negative impact of strong sparsity. Additionally, LPMSS employs the random negative sampling to alleviate the impact of imbalanced link distributions. We also adopt a Times module to capture the temporal topological changes in mobile social networks to enhance the prediction accuracy. Based on three datasets with different sparsity, extensive experiment results demonstrate that LPMSS can significantly improve AUC values while reducing MAE values, confirming its effectiveness in handling the link prediction in the mobile social networks with strong sparsity.

Wireless power transmission has been widely used to replenish energy for wireless sensor networks, where the energy consumption rate of sensor nodes is usually time varying and indefinite. However, few works have investigated the problem of long-term charging scheduling with random variable. This article designs an optimization model for the long-term scheduling of chargers to maximize the time-averaged charging utility while ensuring its time-averaged constraints of budget and response rate. The Lyapunov optimization technique is adopted to transform the stochastic optimization problem into a deterministic optimization problem, which remains NP-hard. Thus, an approximation algorithm following greedy approach is proposed to solve the deterministic optimization problem. We further provide the theoretical analysis of feasibility and performance guarantee of the proposed scheduling algorithm. The simulation results show that our algorithm outperforms three comparison algorithms by 6.53%, 20.04%, and 19.97% in terms of time-averaged charging utility, as well as by 11.25%, 4.42%, and 3.73% in terms of time-averaged response rate on average.

As a complement of public transportation system in modern cities, taxis can provide flexible transportation services for citizens. To increase the profits of taxis, proper cruising routes should be recommended to help vacant taxis find passengers faster. Besides, the cruising tendencies of taxi drivers differ due to the personal cognitions. To this end, we propose a hybrid learning framework-based profit-maximizing personalized route (PMPR) recommendation method for vacant taxis (HL-PPRRM) to obtain and recommend the PMPRs for vacant taxis. HL-PPRRM consists of the local learning on vacant taxis and the global learning on cloud server. Specifically, vacant taxis locally learn the historical cruising routes to predict the personalized routes, while the cloud server globally learns the occupied records to predict the future taxi demand. The obtained PMPRs tally with the cruising tendencies of taxi drivers roughly, and the occupied durations of taxis can be effectively prolonged when vacant taxis cruise along PMPRs. Extensive simulations and comparisons demonstrate the superior performance of our proposed HL-PPRRM, i.e., with the hybrid learning framework, the profits of taxis can be significantly increased, and the average displacement error of PMPRs is very small.

Visible atmospheric particles (such as raindrops, haze particles, and snow particles) could mask the image pixels in the form of some noise, which makes many computer vision tasks difficult to implement. At present, most existing neural network models for snow removal exploit the principle of snow generations, and they remove the snow particles through extracting the snow masks and chromatic aberration maps from the snowy images. Typically, the original image features and the blurred image features are not jointly considered. In this paper, according to the transparency and size of snow particles, also the residual masks between snowy images and snow-free images, we propose a Cascaded Snow Removal Network (CSR-Net). Moreover, the proposed model adopts the collaborative training strategy to accelerate the training process, effectively harnessing the computational resources across multiple devices. Besides, the model parallelism ensures the collaborative training among devices, thereby improving the overall model performance. Specifically, our proposed method can effectively enhance the Peak Signal to Noise Ratio (PSNR) and Structural SIMilarity (SSIM).

Abstract Public transportation system is one of the most effective ways to conserve energy and reduce carbon emissions. However, the traditional public transportation system does not provide customized service and cannot guarantee the arrival time to destination. To address these issues, we formulate the minimum shared bus scheduling problem to minimize the number of shared buses such that all orders can be completed under constraints of deadlines and capacity of shared bus. We propose the approximation algorithms, S‐MBSA for the shared bus with strong endurance and E‐MBSA for the large‐scale order scenario, to solve the minimum shared bus scheduling problem. We further formulate the constrained maximum revenue shared bus scheduling problem to maximize the revenue under the limited number of shared buses, and propose an approximation algorithm, CMRBSA, to find the shared bus route schedules. Through the extensive simulations, we demonstrate the significant superiority of S‐MBSA and E‐MBSA in terms of number of shared buses. Furthermore, CMRBSA outperforms the benchmark algorithms significantly in terms of revenue.

At present, mobile charging stations (MCSs) are taken as an important complement of fixed charging stations. Currently, the strategy of MCSs is to move towards the electric vehicles to be charged (EVCs) only after being requested. To shorten the charging delay of EVCs and enhance the proportion of charged EVCs, idle MCSs should actively move to the areas with large potential charging demand rather than remaining stationary. The distribution of idle MCSs in different areas should be taken into account to prevent excessive idle MCSs from moving into the same areas simultaneously. To this end, we introduce the concept of charging demand force to depict the potential charging demand of EVCs, and then propose the Placement Strategy for Idle Mobile Charging Stations (PS-IMCS). In PS-IMCS, each idle MCS can measure the potential charging demand in neighboring areas through obtaining the resultant force composed of attraction force and repulsion force, and an MDP model is specially designed to make placement decisions for idle MCSs. Extensive simulations and comparisons demonstrate the performance superiority of PS-IMCS, i.e., the charging delay of EVCs can be significantly shortened, and the proportion of charged EVCs can be effectively enhanced.

Mobile charging stations (MCSs) have become an indispensable complement of fixed charging stations. In the regions where fixed charging stations are sparsely deployed or even absent, the main concern is that how to properly schedule MCSs to charge the electric vehicles with insufficient electricity (EVCs). In this paper, we focus on the scheduling of idle MCSs and pending EVCs. To increase the charging revenue of MCSs and enhance the proportion of successfully charged EVCs, we schedule idle MCSs to proactively track some EVCs with potential charging demand, and schedule pending EVCs to approach some busy MCSs for potential charging opportunities. To this end, a Scheduling Approach based on Multi-Agent Deep Reinforcement Learning (SA-MADRL) is proposed to train the scheduling models for agents (idle MCSs and pending EVCs). In SA-MADRL, the agents obtain the local observations to make the scheduling decisions. Both idle MCSs and pending EVCs can independently make the scheduling decisions, and thus SA-MADRL can realize the fully distributed scheduling and has a good scalability. Extensive simulations and comparisons demonstrate the performance superiority of SA-MADRL, i.e., the charging revenue of MCSs can be significantly increased, and the proportion of successfully charged EVCs can be effectively enhanced.

As the gasoline taxis are gradually restricted due to the increased environmental awareness, electric taxis (E-taxis) have become a more environmentally friendly choice to provide the transportation service. When some E-taxis are vacant, they typically cruise along roads without any specific destinations, and two major concerns should be considered for vacant E-taxis: In order to increase the business profits of E-taxis, it is vital to recommend the profitable cruising routes along which vacant E-taxis could pick up passengers as early as possible and earn more profits. Besides, the residual electricity of E-taxis is continuously consumed on travels, and E-taxis must be timely charged before their residual electricity is exhausted (i.e. the breakdowns of E-taxis). Thus, the cruising route recommendation for vacant E-taxis should take into account both passenger-carrying demand and charging demand, and the carrying-charging demand balance should be properly made. To this end, we propose a cruising Route Recommendation Method based on Carrying-charging Demand Balance (RRM-CDB) for vacant E-taxis. The passenger-carrying demand and charging demand are first formulated to reflect their changes and interrelationships, and the historical cruising trajectories of vacant E-taxis (with the two types of demand) are locally learned to recommend the future cruising routes, because the historical cruising trajectories contain the distribution of taxi demand of passengers and the trend of vacant E-taxis gradually approaching the charging stations with the decrease of residual electricity. Particularly, in RRM-CDB each vacant E-taxi trains a local learning model in a distributed manner, thus significantly reducing the computational complexity of cruising route recommendation. Extensive simulations and comparisons demonstrate that RRM-CDB can help to increase the business profits of E-taxis and avoid the breakdowns of E-taxis as much as possible.

Taxis can provide convenient and flexible transportation services for citizens. The proper cruising routes should be recommended to vacant taxis, so as to help them to pick up passengers as early as possible, and thus increase their business profits. To this end, we propose a Cloud-edge-end Collaboration Framework for the Cruising Route Recommendation of vacant taxis (CCF-CRR). In CCF-CRR, each vacant taxi trains a local model based on its historical cruising route segments, and the local model parameters of the vacant taxis in the same region are periodically uploaded to an edge server for parameter aggregation. Then, the aggregated model parameters are released by the edge server to vacant taxis for their use. In addition, the future waiting time of passengers is predicted by the edge servers in different regions and is uploaded to the cloud server, and then the cloud server can measure the potential taxi demand in regions and dispatch vacant taxis among regions to achieve the taxi demand-supply equilibrium. Extensive simulations and comparisons demonstrate the superior performance of our proposed CCF-CRR, i.e., with the cloud-edge-end collaboration framework, the business profits of taxis can be significantly increased, and the pick-up distance of taxis can be largely shortened.

In Internet of Electric Vehicles (IoEV), mobile charging stations (MCSs) have been deployed to complement fixed charging stations. Typically, MCSs are assigned to charge the electric vehicles with insufficient electricity which have made charging requests (termed IEVs). Moreover, there are some electric vehicles with insufficient electricity which have not made charging requests (termed quasi-IEVs). If idle MCSs are allowed to actively track quasi-IEVs according to their potential charging demand, then more IEVs could be promptly charged, and thus the charging profits of MCSs could be increased. However, due to the private ownership of electric vehicles, some private information cannot be provided in the potential charging demand of quasi-IEVs (e.g., the destinations and residual electricity), making the potential charging profits of idle MCSs hard to be evaluated, and thereby the proper assignments of idle MCSs are difficult to decide. To this end, we introduce the profit-maximizing heat maps to depict the potential charging demand of quasi-IEVs and evaluate the potential charging profits of idle MCSs. A profit-maximizing heat map remarks the positions around quasi-IEVs and displays them as continuous areas. Specifically, the different shades of colours are used to distinguish the quantities of potential charging profits of idle MCSs, and the sizes of coloured areas are used to indicate the possibility of quasi-IEVs passing through these positions. In this paper, we propose a Profit-Maximizing Assignment Strategy of Idle MCSs (PMASIM) to properly assign the idle MCSs to charge IEVs at selected charging positions, or track some quasi-IEVs according to the profit-maximizing heat maps. Extensive simulations and comparisons demonstrate the superior performance of PMASIM, i.e., with the profit-maximizing heat maps, the charging profits of MCSs are increased, and the proportion of charged IEVs is enhanced as well.

Wireless Power Transfer (WPT) technology has been developed rapidly in recent years. The cooperative charging model and corresponding scheduling methods have been proposed to save the charging cost in paid charging service. However, the state-of-the-art methods ignore the spatial occupation issue of rechargeable devices. Moreover, the cooperative charging scheduling in directional wireless charging has not been studied yet. This paper studies the cooperative scheduling for directional wireless charging with spatial occupation. We formulate the Cooperative Charging Scheduling with Spatial occupation (CCSS) problem of Mobile Rechargeable Sensor Devices (MRSDs) for optimizing the total cost of whole charging system. We first investigate the properties of optimal arrangement of MRSDs in charging group and calculate the tight intervals of charging angles of MRSDs. We show that it is sufficient to bound the error by conducting angle discretization for only two MRSDs in each charging group. Then, a$(\ln n+1)(1+\varepsilon)$-approximation algorithm of the CCSS problem is proposed based on greedy approach, where$n$is the number of MRSDs, and$\varepsilon$is the discretization error. The results of extensive simulations and field experiments demonstrate that our algorithm can reduce at most 42.5% total cost comparing with the benchmark algorithms.

Wireless Power Transmission (WPT) has been widely used to replenish energy for wireless rechargeable sensor networks. This paper concerns the fundamental issue of robust fault-tolerant placement of wireless chargers for directional charging. Following the general directional charging model, we formulate theCharger Placement for Robust Coverage (CPRC)problem, which has continuous and infinite constraints, for resisting the wireless charger failure. We transform the problem to the equivalent integer program problem without performance loss by area partition and dominating strategy extraction. We show that the greedy algorithm achieves the logarithmic approximation ratio. We further formulate theCharger Placement for Robust Utility (CPRU)problem for resisting the sensor node failure. This problem also has continuous and infinite constraints. We transform the problem to the combinational optimization problem with finite strategy space through the techniques of charging power approximation, area discretization and dominating strategy extraction. We present the algorithm, which utilizes the combination of binary search and greedy algorithm, to solve theCPRUproblem. We conduct both simulations and field experiments to validate our theoretical results. The simulation results show that the proposed algorithms forCPRCandCPRUcan outperform comparison algorithms by at least 17.48% and 21.15%, respectively.

Due to the low cost and agile service provision, mobile charging stations (MCSs) have been deployed to complement fixed charging stations (FCSs). In the Internet of Electric Vehicles (IoEV) with MCSs, a major concern is to enhance the charging efficiency of MCSs. The charging efficiency of MCSs can be improved by prolonging the charging durations of MCSs, i.e., MCSs should undertake the charging tasks as more as possible, which can increase the charging profits of MCSs and reduce the charging expenses of IEVs (EVs with insufficient electricity). Besides, EVs and MCSs are selfish in terms of charging expenses and charging profits, respectively. In this article, we propose a Bayesian game based Bidding Scheme for Mobile Charging enabled Electric Vehicles (BBS-MCEV). In BBS-MCEV, each IEV first calculates the maximum charging price (MCP) according to the potential expense if charged by nearby FCSs, and then the optimal charging price (OCP) is determined by the Bayesian game model. Each MCS accepts the charging request with the largest charging profit. Extensive simulations and comparisons demonstrate the superior performance of our proposed BBS-MCEV, i.e., with the Bayesian game model, IEVs can rationally bid for the mobile charging services from MCSs, and thus BBS-MCEV can increase the charging profits of MCSs and reduce the charging expenses of IEVs effectively. Besides, a proper tradeoff between the charging profits of MCSs and the charging expenses of IEVs can be achieved.

Abstract Delay tolerant networks (DTNs) refer to a kind of novel wireless mobile network, where there is no constant end‐to‐end connection between network nodes due to frequent movement, sparse distribution and limited communication range of nodes. Instead of the traditional store‐forward routing strategy, in DTNs, the new store‐carry‐forward routing strategy is adopted for data transmission. Therefore, how to select the best next‐hop node among network nodes is the main challenge of the routing in DTNs. To this end, here, a k ‐step double Q‐learning routing (K‐DQLR) algorithm is proposed, which integrates the multi‐step and double Q‐learning algorithms to make an unbiased, accurate and efficient routing decision in DTNs. Besides, a new dynamic reward mechanism is proposed, which combines the number of routing hops and the node centrality to adopt the dynamic network environment of DTNs. The simulation results show that K‐DQLR can significantly increase the delivery ratio while reducing the delivery delay and overhead compared with the related state‐of‐the‐art routing protocols of DTNs.

The electric vehicles (EVs), as promising components of sustainable and eco-friendly transportation systems, are being widely adopted to reduce the consumption of fossil fuel and the pollution of environments. EVs are usually equipped with wireless communication modules to support the vehicle to vehicle (V2V) communications, and thus an electric vehicular network (EVN) is constituted. However, the electric energy of EVs is extremely limited, and some EVs are possible to exhaust their electric energy before reaching the destinations. More seriously, when the EVs travel into the zones without any charging stations, they cannot be timely charged. With the rapid developments of wireless charging technologies (such as magnetic resonance) and graphene supercapacitor technologies, the reciprocal charges between EVs become feasible, i.e., the EVs with insufficient energy (IEVs) can be charged by the EVs with surplus energy (SEVs). In this paper, the strategies of selecting the local-optimal SEVs for IEVs and rescheduling their travel routes are investigated, and a distributed Reciprocal Charging Mechanism (RCM) is proposed. Both mechanism analysis and simulation results demonstrate the performance superiority of RCM. Specifically, with the proposed reciprocal charging mechanism, IEVs can be charged by SEVs in a charging-station-absent zone, and the electric energy consumption can be approximatively minimized.

The multi-hop wireless charging technology can largely extend the charging service range of chargers, thus has promising prospect in sustainable energy replenishment for wireless rechargeable sensor network. This paper proposes a new cost criterion, termed comprehensive cost consisting of energy cost and deployment cost, to measure the actual expenditure of wireless charging. We present a multi-hop wireless charging model and formulate the problem of minimizing the comprehensive cost such that the energy demand of all sensor nodes can be fulfilled by the energy capacitated chargers. We propose a (ln n+1)-approximation algorithm for the optimization problem, where n is the number of sensor nodes. Then, we propose a straightforward cost sharing mechanism, which ensures that no subset of sensor nodes can benefit by breaking away from the current charging tree for any fixed charger position, to realize the paid charging service of multi-hop wireless charging. Furthermore, to keep the magnetic fields of transmitters from the interfering, the conflict avoidance schemes are proposed in both central and distributed situations. Finally, we discuss the distributed scheme for minimizing the comprehensive cost without support of central server. Through extensive simulations, we demonstrate the significant superiority of the proposed algorithms in terms of comprehensive cost.

Mobile crowdsourcing has become an efficient paradigm for performing large scale tasks. The incentive mechanism is important for the mobile crowdsourcing system to stimulate participants, and to achieve good service quality. In this paper, we focus on solving the insufficient participation problem for the budget constrained online crowdsourcing system. We present a two-tiered social crowdsourcing architecture, which can enable the selected registered users to recruit their social neighbors by diffusing the tasks to their social circles. We present three system models for two-tiered social crowdsourcing system based on the arrival modes of registered users and social neighbors: offline model, semi-online model, and full-online model. We consider the tasks are associated with different end times. We present an incentive mechanism for each of three system models. Through both rigorous theoretical analysis and extensive simulations, we demonstrate that the proposed incentive mechanisms achieve computational efficiency, individual rationality, budget feasibility, cost truthfulness, and time truthfulness. We further show that our incentive mechanisms for semi-online model and full-online model can obtain averagely 51.1$\%$and 39.7$\%$value of approximate optimal untruthful offline algorithm, respectively.

The multi-hop wireless charging technology has attracted a lot of attention, as it largely extends the charging range of chargers. Different from the existing work with single cost optimization, the objective of this article is to optimize the comprehensive cost, which is the combination of energy cost and deployment cost. We decompose the target problem into two sub-problems. The first sub-problem aims to minimize the deployment cost with energy capacity constraints. The proposed algorithm follows the greedy strategy, where the subset of sensor nodes for any charger is determined by finding the capacitated minimum spanning tree. The second sub-problem, which aims to maximize the reduction of comprehensive cost by adding chargers to the solution of the first sub-problem, is proved to be an unconstrained submodular set function maximization problem and can be solved by a 1/2-approximation randomized linear time algorithm for its equivalent problem. Through extensive simulations, we demonstrate that the proposed solution can reduce the comprehensive cost by 57.55% comparing with the benchmark algorithms.

The movement of a cruising ET (which do not pick up passengers) is quite sophisticated, and it depends on the driver’s subconscious movement tendency, driving habits, and historical pick-up experiences. This paper exploits the links between ETs and charging stations according to historical trajectories, and proposes a Charging station Recommendation Algorithm based on Link Prediction (CRA-LP) to recommend the optimal charging stations for cruising ETs. Specially, in CRA-LP, the Adamic-Adar index with a time decay effect is given to measure the similarities between ETs and charging stations. Simulations results show that CRA-LP achieves preferable results in terms of some metrics, such as Area Under Curve and Average Extra Movement, which also indicates that CRA-LP can predict the future links between cruising ETs and charging stations accurately, and then recommend the proper charging stations to cruising ETs to minimize the extra movements.

Opportunistic Underwater Sensor Network (OUSN) comprised of Unmanned Underwater Vehicles (UUV) has been deployed for the surveillance of various underwater events, and in some military environments an OUSN could be invaded by some underwater spy-robots termed eavesdroppers. Typically, the eavesdroppers move around OUSN nodes and eavesdrop on their communication channels silently. Thus, the eavesdroppers are hard to be perceived by the nodes, which implies that the nodes are blind to the eavesdroppers. The nodes probably disseminate the held data messages while they are blithely unaware of the adjacent eavesdroppers, and the eavesdroppers could capture these data messages. To reduce the theft ratio of data messages and guarantee the required delivery ratio in a storage-limited OUSN, a propagation model of data messages is formulated to analyse the proportion variation of message holders. Based on this model, the reduction of theft ratio is specially investigated to obtain the appropriate disseminating probabilities, storing probabilities, and discarding probabilities of the nodes with different indegrees. In the proposed Blind Message Dissemination Method (BMDM), the data messages are disseminated, stored, and discarded according to these probabilities. Simulation results demonstrate that BMDM can reduce the theft ratio and guarantee the required delivery ratio in a storage-limited OUSN effectively.

Task graphs are widely used to represent data-intensive applications. To efficiently execute these applications on heterogeneous systems, each task must be properly scheduled on the processors of the system. The NP-completeness of the task scheduling problem has motivated researchers to propose various heuristic methods. Recently, Quality of Service (QoS) aware scheduling is becoming an active research area in heterogeneous systems because the end-user has different QoS requirements. Generally, time and cost are the most relevant user concerns. However, it is challenging to find a feasible scheduling plan which minimizes the total execution time of the user’s application (makespan) while satisfying both budget and deadline constraints. In this paper, we present a novel heuristic algorithm called Budget-Deadline-Aware-Scheduling (BUDA) that addresses task graphs scheduling under budget and deadline constraints in heterogeneous systems. The novelty of the BUDA algorithm is based on a Heterogeneous Time-Cost Matrix (HTCM) that is used to prioritize tasks and for processor selection. In addition, we introduce a new Heterogeneous Time-Cost Trade-off factor (HTCT) that tries to adjust the time and cost for the current task among all processors. The experiments based on randomly generated graphs and real-world applications graphs show that the BUDA algorithm outperforms the state-of-the-art algorithms in terms of makespan, time efficiency, and success rate.

Abstract Opportunistic underwater sensor network (OUSN) is deployed for various underwater applications, such as underwater creatures tracking and tactical surveillance. Particularly, the OUSN in military applications may be invaded by some underwater spy‐robots termed eavesdroppers. The eavesdroppers could move around some OUSN nodes and eavesdrop on their communication channels silently, and these eavesdropping actions are difficult to be perceived by OUSN nodes. To reduce the theft ratio of data messages and guarantee the required delivery ratio of data messages, we conceive the idea that each data message is encoded into several message pieces, and then the message pieces are disseminated to sink node individually. Besides, a lightweight encryption method is adopted to encrypt the message pieces before the dissemination. Such mechanism can protect the data messages from being stolen by eavesdroppers effectively. In this article, we propose a message piece dissemination approach (MPDA) for the OUSN invaded by some underwater spy‐robots. In MPDA, OUSN nodes disseminate the held message pieces to some selected neighboring nodes at each time slot, and a data message is considered to be delivered when all pieces of this data message have been delivered to the sink node. Extensive simulations and comparisons demonstrate the preferable performance of MPDA, that is, MPDA can reduce the theft ratio of data messages and guarantee the required delivery ratio of data messages.

Opportunistic Underwater Sensor Networks (OUSNs) are deployed for various underwater applications, such as underwater creature tracking and tactical surveillance. In an OUSN invaded by some eavesdroppers, the data messages disseminated by sensor nodes are probably stolen (captured and cracked) by the eavesdroppers. The data messages are disseminated through acoustic waves which could be altered by the environmental noises, i.e., the acoustic waves containing data messages could be superimposed by the environmental noises. To protect the data messages from being stolen by eavesdroppers and guarantee the required delivery ratio of data messages, we propose a Noise-based-protection Message Dissemination Method (NMDM). In NMDM, the acoustic waves containing data messages are superposed by the environmental noises and converted into some pseudo data messages. The environmental noises around source nodes are identified, encoded, and encrypted into some noise messages. Then, the pseudo data messages and noise messages are individually disseminated to the sink node. Such mechanism makes the eavesdroppers difficult to steal the data messages. Besides, the required delivery ratio of data messages is achieved by measuring the similarities between the nodes and the sink node, i.e., the pseudo data messages and noise messages are preferentially disseminated to the nodes with larger similarities to the sink node. Finally, simulation results demonstrate the superior performance of NMDM. NMDM can reduce the theft ratio of data messages and guarantee the required delivery ratio of data messages effectively.

Electric vehicles (EVs), as promising components of the sustainable and eco-friendly transportation systems, are being widely adopted to reduce the consumption of fossil fuel and pollution of environments. EVs are usually equipped with wireless modules to support the vehicle to vehicle communications, by which an electric vehicular network (EVN) is formed. In EVN, some EVs are with insufficient battery energy and may exhaust the battery energy before arriving at their destinations, and these EVs are referred to as IEVs. More seriously, IEVs probably cannot find any fixed charging facilities nearby. With the development of mobile charging technology, some movable charging stations (MCSs) are deployed into EVN, and MCSs can actively navigate to charge IEVs. In this paper, an assignment rescheduling mechanism of movable charging stations (ARMM) is proposed, where the MCS assignments are dynamically rescheduled. In ARMM, in order to reduce the charging expenses of IEVs and enhance the proportion of charged IEVs, the assigned IEVs of some MCSs could be switched to other MCSs, while the charging positions of MCSs are selected by minimizing the charging expenses of IEVs and are dynamically altered. Besides, the incentives of assigned IEVs to reduce the charging expenses of unassigned IEVs are proven. Simulation results demonstrate the preferable performance of ARMM, i.e. ARMM can reduce the charging expenses of IEVs and enhance the proportion of charged IEVs effectively.

In Internet of Electric Vehicles (IoEV), mobile charging stations (MCSs) can be deployed to complement fixed charging stations. Currently, the strategy of MCSs is to move towards the EVs with insufficient energy (IEVs) only after being requested, which is not efficient. However, similar to online car-hailing services, more IEVs could be charged and the charging expenses could be reduced if idle MCSs can actively move towards the potential charging positions. In this paper, the problem of placements of idle MCSs in an IoEV is investigated in order to enhance the proportion of charged IEVs and reduce the charging expenses of IEVs. To this end, we propose a Federated Learning based Placement Decision Method of Idle MCSs (FL-PDMIM) to help the idle MCSs to predict the future charging positions, by exploiting the historical routes of MCSs which contain rich information regarding the charging demand of IEVs. In the proposed framework, the historical routes are trained locally by each MCS, and then the local model parameters and charging records are periodically uploaded to an edge server for a global parameter aggregation. Then, idle MCSs decide their placements according to the predicted charging positions (potential charging positions). The training time can be largely shortened, because the distributed learning on each MCS is executed in parallel. Extensive simulations and comparisons demonstrate the performance superiority of FL-PDMIM. Specifically, with the proposed federated learning-based predictions, the waiting time of IEVs to be served can be significantly shortened, and FL-PDMIM enhances the proportion of charged IEVs and reduces the charging expenses of IEVs effectively.

Opportunistic underwater sensor networks (OUSNs) are deployed for various underwater applications, such as underwater creatures tracking and tactical surveillance. The data dissemination in OUSNs differs significantly from those in terrestrial wireless sensor networks or delay-tolerant networks, due to the signal irregularity in underwater communications and the limited storage capacity of the nodes in OUSNs. To alleviate the storage overflows on nodes and make room for the newly arriving data packets, some stored data packets ought to be actively discarded by nodes. This research begins with the construction of a differential equation set to describe the propagation process of data packets in OUSNs, and the storing-discarding equilibrium is investigated such that each data packet is expected to propagate and disappear during the allowable dissemination time slots. After that, the optimal storing probabilities and discarding probabilities are obtained for the nodes with different in-degrees to maximize the delivery ratio of data packets. Then, we propose an Adaptive Data Dissemination Algorithm (ADDA) for the storage-limited OUSNs with signal irregularity, where at each time slot the newly arriving data packets are stored and the stored data packets are discarded by nodes according to the obtained storing probabilities and discarding probabilities, respectively. Simulation results demonstrate the excellent performance of ADDA, showing that it can enhance the delivery ratio of data packets and reduce the number of storage overflows.

Abstract The technique of vehicle license plate recognition can recognize and count the vehicles automatically, and thus many applications regarding the vehicles are greatly facilitated. However, the recognitions of vehicle license plates are extremely difficult especially in some fog‐haze environments because the fog and haze blur the boundaries and characters of license plates significantly, which makes the license plates hard to be detected or recognised. To this end, this paper proposes a vehicle License Plate Recognition method for Fog‐Haze environments (LPRFH). In LPRFH, a dark channel prior algorithm based on the local estimation of atmospheric light value is applied to dehaze the blurred images preliminarily. Then, the images are further dehazed, and the license plate regions are detected through a Joint Further‐dehazing and Region‐extracting Model on basis of an object detection convolution neural network. Finally, the image super‐resolution is accomplished with a convolution‐enhanced super‐resolution convolutional neural network, and hence the characters of license plates can be recognised successfully. Extensive experiments have been conducted, and the results indicate that LPRFH can recognise the license plates accurately even in some severe fog‐haze environments.

Abstract As the basis of image processing, single image deraining has always been a significant and challenging issue. Due to the lack of real rainy images and corresponding clean images, most deraining networks are trained by synthetic datasets, which makes the output images unsatisfactory in real applications. Besides, note that a heavy rainfall is typically accompanied with some fog. Although some deraining networks have been proposed to remove the rain streaks in the rainy images, the output images may still be blurred due to the accompanied fog. In this paper, these problems existing in single image deraining is comprehensively considered, and propose a Cycle‐Derain network based on an unsupervised attention‐guided mechanism. Specifically, the Cycle‐Derain network takes advantage of generative adversarial networks with two mappings and the cycle consistency loss to train both unpaired rainy images and rain‐free images. Moreover, it introduces an unsupervised attention‐guided mechanism and exploits the loop‐search positioning algorithm to deal with the details of rain and fog in images. Extensive experiments have been carried out, and the results show that the proposed Cycle‐Derain network is preferable compared with other deraining networks, especially in term of rainy image restoration.

Underwater acoustic sensor networks (UASNs) have attracted considerable attention due to the increasing demands on the Internet of Underwater Things. The monitored underwater data are disseminated by anchored nodes in UASNs through multihop transmissions. In the military applications, some underwater spy robots are probably dispatched by the enemy to invade the UASNs, and these underwater spy robots seek to steal the data messages disseminated by anchored nodes. In this article, the autonomous underwater vehicles (AUVs) are taken as the ferries for the data message dissemination of some vulnerable anchored nodes, so as to reduce the risk of data messages being stolen by the spy robots. To this end, an index dissemination vulnerability is first introduced to measure the risk of data messages being stolen around anchored nodes, and the network topology is specially investigated to ameliorate the dissemination vulnerabilities of anchored nodes. An AUV-aided message ferry approach (AMFA) is proposed for the UASNs invaded by underwater spy robots. In AMFA, the anchored nodes set the initial communication ranges according to a power law distribution, and then, the communication ranges of anchored nodes are updated to achieve the required topology connectivity. Especially, some AUVs are assigned to ferry the data messages of the anchored nodes with the largest dissemination vulnerabilities. The simulation results demonstrate the preferable performance of AMFA, i.e., AMFA ameliorates the dissemination vulnerabilities of anchored nodes considerably while the required topology connectivity can be guaranteed.

Cloud technology has brought great convenience to enterprises as well as customers. System logs record notable events and are becoming valuable resources to track and investigate system status. Detecting anomaly from logs as fast as possible can improve the quality of service significantly. Although many machine learning algorithms (e.g., SVM, Logistic Regression) have high detection accuracy, we find that they assume data are clean and might have high training time. Facing these challenges, in this paper, we propose Robust Online Evolving Anomaly Detection (ROEAD) framework which adopts Robust Feature Extractor (RFE) to remove the effects of noise and Online Evolving Anomaly Detection (OEAD) to dynamic update parameters. We propose Online Evolving SVM (OES) algorithm as the example of online anomaly detection methods. We analyze the performance of OES in theory and prove the performance difference between OES and the best hypothesis tends to zero as time goes infinity. We compare the performance of ROEAD against state-of-the-art anomaly detection algorithms using public log datasets. The results demonstrate that ROEAD is able to remove the effects of noise and OES can improve the detection accuracy by more than 40%.

Delay‐tolerant networks (DTNs) are wireless mobile networks, which suffer from frequent disruption, high latency, and lack of a complete path from source to destination. The intermittent connectivity in DTNs makes it difficult to efficiently deliver messages. Research results have shown that the routing protocol based on reinforcement learning can achieve a reasonable balance between routing performance and cost. However, due to the complexity, dynamics, and uncertainty of the characteristics of nodes in DTNs, providing a reliable multihop routing in DTNs is still a particular challenge. In this paper, we propose a Fuzzy‐logic‐based Double Q‐Learning Routing (FDQLR) protocol that can learn the optimal route by combining fuzzy logic with the Double Q‐Learning algorithm. In this protocol, a fuzzy dynamic reward mechanism is proposed, and it uses fuzzy logic to comprehensively evaluate the characteristics of nodes including node activity, contact interval, and movement speed. Furthermore, a hot zone drop mechanism and a drop mechanism are proposed, which can improve the efficiency of message forwarding and buffer management of the node. The simulation results show that the fuzzy logic can improve the performance of the FDQLR protocol in terms of delivery ratio, delivery delay, and overhead. In particular, compared with other related routing protocols of DTNs, the FDQLR protocol can achieve the highest delivery ratio and the lowest overhead.

Summary Sketch is a compact data structure used to summarize data streams. It is widely used in the measurement of network traffic, and its accuracy is higher than traditional methods. Currently, there are some typical sketches: Count‐Min Sketch, CU Sketch, and Count Sketch. According to the characteristics of network traffic, we propose a new sketch framework called Self‐Adaption Sketch, which is combined Sketch with Bloom Filter. In the framework, the sketch is created dynamically and the memory space is adjusted timely according to the network traffic by using the concept carrying. Our experiment results showed that the space utilization and accuracy are significantly improved while the throughput of self‐adaption sketch is maintained at a relatively good level.

Delay tolerant networks (DTNs) are wireless mobile networks, where the nodes are sparse and end-to-end connectivity is rare. The intermittent connectivity in DTNs makes it challenging to efficiently deliver messages. Research results have shown that the routing protocol based on reinforcement learning can achieve a reasonable balance between routing performance and cost. However, how to predict the next hop of messages more accurately is still open. In this paper, Double Q-Learning Routing (DQLR) protocol is proposed, which investigates the routing selection of the next hop in a distributed manner and solves the overestimation problem by Double Q-Learning algorithm. Further, the intermediate value and dynamic reward mechanisms are proposed to adapt node mobility and network topology change, which improve the network performance. The simulation results show that DQLR protocol can increase the delivery ratio with a low overhead.

As an important issue in the trajectory mining task, the trajectory clustering technique has attracted lots of the attention in the field of data mining. Trajectory clustering technique identifies the similar trajectories (or trajectory segments) and classifies them into the several clusters which can reveal the potential movement behaviors of nodes. At present, most of the existing trajectory clustering methods focus on some spatial properties of trajectories (such as geographic locations, movement directions), while the spatial-temporal properties (especially the combination of spatial distances and semantic distances) are ignored, and thus some vital information regarding the movement behaviors of nodes is probably lost in the trajectory clustering results. In this paper, we propose a Joint Spatial-Temporal Trajectory Clustering Method (JSTTCM), where some spatialtemporal properties of the trajectories are exploited to cluster the trajectory segments. Finally, the number of clusters and the silhouette coefficient are observed through simulations, and the results show that JSTTCM can cluster the trajectory segments appropriately.

The Underwater Sensor Network (USN) has great advantages in marine environmental monitoring. When collecting the perception information of sensor nodes, the mobile node can effectively compensate for the shortcomings of traditional multi-hop transmission modes. However, the complex marine environment causes the location uncertainty of nodes. Therefore, a path planning method based on the location uncertainty of water surface nodes in USN is proposed in this paper. Firstly, the structure of the USN based on the mooring model is introduced and the problem model of path planning is proposed. Secondly, the inevitable communication circle is obtained by analyzing the deviation range and communication range of water surface nodes. Thirdly, Convex Hull algorithm is used to plan the path in accordance with the inevitable communication circle of water surface nodes. Finally, simulation results show that the proposed method can obtain a shorter path under the premise of ensuring the completion of information collection.

The opportunistic mobile sensor network has been extensively applied in various public safety applications such as the fire rescue and earthquake rescue, since it can provide a surveillance range with an inexpensive cost and avoid the dangers of humans staying in risk zones. However, due to some environmental events such as building structure damage, airflow push, and fire explosions, the sensor nodes sprinkled into the fire-rescue scenario may be kept moving. Thus, the contacts between nodes become momentary, and the data packets cannot be forwarded along stable communication paths. To this end, the opportunistic forwarding manner is adopted in the fire-rescue scenario to enable the data packets to be transferred to the rescue control center (RCC) through some discrete hops. The contributions of this paper are threefold. First, the nodes in the fire-rescue scenario are carefully investigated and classified into four types: small-range mobile nodes (SRNs), large-range mobile nodes (LRNs), firefighter nodes (FNs), and robot nodes (RNs). Second, we formulate the data forwarding problem, and the optimal proportions of SRNs, LRNs, and FNs in data holders are mathematically analyzed to obtain the maximum delivery ratio. Third, a data forwarding approach for fire-rescue scenario (DFAFR) is proposed. In DFAFR, the optimal proportions of SRNs, LRNs, and FNs in data holders are maintained as far as possible through selecting different types of data holder candidates, and then the new data holders are determined from these data holder candidates and the adjacent RNs on basis of their expected delivery delay. Finally, the performance of DFAFR is analyzed through simulations of the fire-rescue scenario, and the results indicate that DFAFR can enhance the delivery ratio and shorten the delivery delay while the forwarding overhead is restricted.

Trajectory data mining has become an increasing concern in the location-based applications, and the trajectory partition is taken as the primary procedure of trajectory data mining. The amount of movement trajectories of nodes is typically very large, and the trajectory shapes are extremely diverse, which makes the trajectory partition a vital issue to the trajectory data mining results. In this work, the movement behaviors of nodes are analyzed from the aspects of moving speeds, stop points, and moving directions, and then a novel Trajectory Partition Method based on combined movement Features (TPMF) is proposed to partition the trajectories. In TPMF, we first extract the change points where the movement speeds of nodes are varied significantly; then, we extract the stop points by detecting the speed variations of nodes; finally, the Douglas-Peucker algorithm is applied to partition the subtrajectories according to the extracted feature points (change points and stop points). Simulations are carried out on the Geolife trajectory dataset, and the simulation results indicate that TPMF can achieve a preferable trade-off between the simplification rate and the trajectory partition error, while the running time is shortened as well.

The opportunistic mobile sensor network has been extensively used in various public safety applications such as the fire-rescue scenario, since it can provide a surveillance range with an inexpensive cost and avoid the dangers of staying in the risk zones to humans. However, the sensor nodes thrown by firefighters in the fire-rescue scenario are easy to move away from current positions due to many environmental factors such as the building structure damages, airflow push or even some explosions. Consequently, the contacts between nodes become scarce and momentary, thereby making the gathered data packets difficult to be forwarded along stable communication paths. Firstly, the mobility patterns of nodes in the fire-rescue scenario are classified into three types: small-range mobile nodes, large-range mobile nodes and firefighter nodes. Then, the optimal proportions of different types of nodes in the data holders are specially investigated mathematically to maximize the delivery ratio. Thus, a data forwarding approach for fire-rescue scenario (DFAFR) is proposed. In DFAFR, each data holder forwards the held data packets to neighbouring nodes independently, and the optimal proportions of data holders are maintained approximatively. Finally, the performance of DFAFR is analyzed through simulation experiments that produce preferable results in the fire-rescue scenario, indicating that DFAFR can improve the delivery ratio and shorten the delivery delay, so that the fire behavior can be reported and processed timely.

In this paper, we investigate the profit maximization problem for the mobile virtual network operator in cognitive radio networks considering the uncertain property of users' spectrum demand. In order to achieve more revenues while simultaneously satisfying the needs of users, the cognitive mobile virtual network operator chooses to dynamically sense the idle spectrum in the licensed band which is more economic, and at the same time leases the spectrum from the spectrum owner which guarantees more stable spectrum resources. However, the fluctuant spectrum demand of users imposes unprecedented challenges on the decision making process. To deal with the uncertain features of the users' demand, a flexible distribution uncertainty model is developed. Particularly, a reference distribution is introduced based on historical data and then a uncertainty set is defined to confine the spectrum demand. The uncertainty model developed allows the actual users' spectrum requirement to fluctuate around the reference distribution. Chance constraint approximations and robust optimization approaches are developed to transform and then solve the optimization problem. Simulation results based on the real-world traces evaluate the performance of the proposed scheme and investigate the parameter impacts on the system utilities. Our research may also help shed some insights on the investment policy making for the mobile virtual network operator.

Delay-tolerant networks (DTNs) are wireless mobile networks where constant end-to-end connections may not exist among nodes. In real-life vehicle DTNs, most nodes have repetitive movement patterns. However, due to the change of time and different activity scenarios, the movement patterns cannot be described consistently with a single model. Considering this issue, the Multi-period Bayesian Network (MBN) is proposed to build multiple prediction models, which intends to predict the regular movement patterns of nodes in the real world. The Bayesian network model is constructed by using several network parameters (e.g. spatial and temporal information at the time of message forwarding) to describe the movement patterns of DTN nodes. Additionally, a novel classification method called Dynamic Multiple-Level Classification (DMLC), is proposed where nodes are classified into multiple levels according to the dynamic parameters. Followed by that, a routing algorithm based on MBN is presented, which can make routing decisions based on the classification results of DMLC. The simulation results show that MBN algorithm and DMLC method can improve the delivery ratio with a minor forwarding overhead.

In Mobile Opportunistic Networks (MONs), due to the node movements and the uncontrollable on/off switches of the carried communication devices, the contacts between nodes may be scarce and momentary, and thus a data packet should be transferred through some discrete hops. To avoid the costly flooding of data packets, the data packets are typically disseminated to some relay nodes selected by data holders. However, the mobility patterns of nodes will become different in different types of regions (such as residential regions, commercial regions, scenery regions, or industrial regions); i.e., the movement directions and movement ranges of nodes are frequently varied when the nodes move among various regions. At present, the issues regarding the region types and region type correlations have not been investigated for the data dissemination in existing works. To this end, we propose a Region Type based Data Dissemination Method (RTDDM) for MONs, which exploits the region type correlations and selects the proper relay nodes through a Markov decision model. To verify the performance of RTDDM, we give some theoretical analysis as well as an elaborated simulation study, the results of which show that RTDDM can improve the delivery ratio and reduce the delivery delay, especially in the applications with various region types.

Delay-tolerant networks (DTNs) are wireless mobile networks that exhibit frequent intermittent connectivity and large transmission delay among nodes. Research results have shown that quota-controlled routing protocols can strike a reasonable balance between routing performance and cost, where quota is a value to control the number of message copies. However, the question of how to set the optimal quota dynamically in order to achieve the lower bound of routing cost is still open. In this paper, we model the optimization of quota control as an extremal functional problem and analyze it by a classic mathematical method called Calculus of Variations (CoV) for the first time. The function of time-variant quota with minimal average number of message copies is obtained in closed form, and an optimal quota control algorithm is proposed under practical routing design considerations. Both the numerical and simulation results show that the proposed model and algorithm are effective and efficient.

Delay-Tolerant Networks (DTNs) are wireless mobile networks, where the nodes are sparse and end-to-end connectivity is rare. Since DTN nodes are mostly energy-limited devices, there is an immediate need to have energy-efficient routing protocols, allowing the network to perform better and function longer. Besides, in the real world, people carrying the nodes form a lot of communities because of similar interests, and they behave with social selfishness. How to improve the energy efficiency in multi-community scenarios has been an important problem. In this paper, we analytically model the performance of epidemic routing protocols in multi-community scenarios with social selfishness considerations using the Ordinary Differential Equations (ODEs). Further, an energy-efficient copy-limit-optimized algorithm based on the Box's complex method for epidemic routing is proposed, which is designed to determine the optimal copy limit in multiple communities, and can improve the energy efficiency effectively. At last, both the numerical and simulation results show that the routing protocol with the proposed algorithm can reduce the energy consumption effectively, and the impact of social selfishness is also analyzed.

Opportunistic underwater sensor networks (OUSNs) are developed for a set of underwater applications, including underwater creatures tracking and tactical surveillance. However, the storage capacity of nodes is sometimes insufficient, especially compared to a wealth of data messages which are generated rapidly in some emergency response applications. Therefore, the network throughput should be taken as one of the primary objectives of message dissemination. To this end, the strategies for message storing, disseminating and discarding are investigated, and a Message Dissemination Approach for Storage-Limited (MDA-SL) OUSNs is proposed hereby. In MDA-SL, the messages are preferred to be disseminated to the nodes with higher speed or larger residual storage. In addition, the newer messages are inclined to be discarded when their holders' storage is full. Furthermore, through simulation analysis, the performance of MDA-SL is proved excellent, which indicates that MDA-SL achieves a satisfactory throughput with the propagation delay being restricted according to application requirements.

Node mobility and end-to-end disconnections in Delay Tolerant Networks (DTNs) greatly weaken the effectiveness of data transmission. Although social-based strategies can be used to deal with the problem, most existing approaches adopt multicopy strategy to forward messages which inevitably add more unnecessary cost. One of the most important issues is the selection of the best intermediate node to forward messages to the destination node. In this paper, we focus on finding a quality metric associated with better relays which is evaluated by Reachable Probability Centrality (RPC) as we proposed. RPC combines the contact matrix and multi-hop forwarding probability based on the weighted social network, thus ensuring an effective relay selection. We also propose a distributed RPC-based routing algorithm, which demonstrates the applicability of our scheme in the decentralized environment of DTNs. Extensive trace-driven simulations show that RPC outperforms other centrality measures and our proposed routing algorithm can significantly reduce the data forwarding cost while having comparable delivery ratio and delay to those of the Epidemic routing.

Underwater wireless sensor networks (UWSNs) have been developed for a set of underwater applications, including the resource exploration, pollution monitoring, tactical surveillance, and so on. However, the complexity and diversity of the underwater environment differentiate it significantly from the terrestrial environment. In particular, the coverage requirements (i.e., coverage degrees and coverage probabilities) at different regions probably differ underwater. Nevertheless, little effort has been made so far on the topology control of UWSNs given the diverse coverage requirements. To this end, this article proposes two algorithms for the diverse coverage problem in UWSNs: (1) Traversal Algorithm for Diverse Coverage (TADC), which adjusts the sensing radii of nodes successively, that is, at each round only one node alters its sensing radius, and (2) Radius Increment Algorithm for Diverse Coverage (RIADC), which sets the sensing radii of nodes incrementally, that is, at each round multiple nodes may increase their sensing radii simultaneously. The performances of TADC and RIADC are analyzed through mathematical analysis and simulations. The results reveal that both TADC and RIADC can achieve the diverse coverage while minimizing the energy consumption. Moreover, TADC and RIADC perform nicely in obtaining optimal sensing radii and reducing message complexity, respectively. Such merits further indicate that TADC and RIADC are suitable for small-scale and large-scale UWSNs, respectively.

Underwater acoustic sensor networks (UASNs) have been developed for a set of underwater applications, including resource exploration, pollution monitoring, and tactical surveillance. Topology control techniques of UASNs are significantly different from those of terrestrial wireless sensor networks, due to the properties of underwater environments and acoustic communications. This research begins with a scale-free network model for calculating edge probability, which is used to generate initial topology randomly. Subsequently, a topology control strategy based on complex network theory (TCSCN) is put forward to construct a double clustering structure, where there are two kinds of cluster-heads to ensure connectivity and coverage, respectively. The performance of TCSCN is analyzed through simulation experiments that indicate a well-constructed topology, where (1,$\xi$)-Coverage and (1,$\zeta$)-Connectivity can be achieved while optimizing energy consumption and propagation delay as much as possible.

Signal irregularity phenomenon affecting network protocols is prone to exhibit in underwater environments, and previous efforts about topology control of underwater sensor networks either have not been made about signal irregularity, or the irregularity models are overly utopian. This paper constructs a more authentic signal irregularity model, which can be degenerated into a variety of special cases easily, and three representative topology control objectives ((KS,β)-Coverage, (KC,α)-Connectivity, and efficient consumption) are concluded. A topology control algorithm for signal irregularity (TCAI) is designed for this topology control problem. The results show the convergence of TCAI and polynomial complexity as well. The performance of the algorithm is also analyzed through simulations that indicate a well-constructed topology, where (KC,β)-Coverage and (KC,α)-Connectivity can be achieved while optimizing energy consumption as much as possible.