Recently, the edge-aided vehicular crowdsensing (EAVC) system has become a promising data collection mode, which utilizes vehicles to collect sensing data under the guidance of edge servers. Participant selection is a fundamental problem in vehicular crowdsensing. The available relevant schemes are unsuitable for newly arrived participants, ignore the differentiated sensing requirements of different areas, and underestimate heterogeneity among participants, which seriously damages the service quality. To handle these problems, this paper proposes an improved reinforcement learning-based online participant selection scheme incorporating coverage quality and participant ability (PSCQA) in EAVC. Coverage quality considering different spatiotemporal partitions is formulated based on the entropy theory to measure coverage uniformity of all areas and the coverage degree of the hotspot areas . Participant ability is designed by combining data quality , movement predictability, and priorities of passing areas to comprehensively measure performance differences among participants. In PSCQA, the coverage quality and ability of the selected participants are optimized through two separate value functions . In particular, participants are dynamically grouped into vehicle clusters based on the similarity of their trajectories to solve the state explosion problem that plagues traditional Q-learning. Simulation results on a real-world dataset demonstrate that the proposed PSCQA outperforms other reinforcement learning-based online participant selection schemes and traditional offline participant selection schemes.

Vehicle re-identification primarily faces two challenges: significant intra-class variations caused by different camera views and subtle inter-class differences between vehicles of the same model. In this paper, we propose a Multi-View Progressive Graph Interaction Embedding Network (MP-GIEN) for vehicle re-identification. First, we design a global semantic extractor. By employing Graph Interaction Units (GI Units) and semantic context, we combine semantic information to facilitate contextual reasoning of image regions, thereby extracting global semantic features of the target. The global loss is then calculated using ID loss and triplet loss. Second, we devise a local view-aware extractor. The U-Net network is utilized to parse the image into four views (front, rear, top, and side). Subsequently, features are aligned through mask average pooling, and a progressive training strategy is adopted to address the learning of fine-grained information. Lastly, we design a feature fusion enhancer. We revise the typical triplet loss to avoid mismatches in local features. By optimizing the local triplet loss and global loss, we learn the embedding of visual features, which not only reduces the intra-instance distance but also enlarges the inter-instance differences. MP-GIEN facilitates the capture of stable discriminative information on vehicles under different views. Experiments conducted on the VeRi-Wild dataset demonstrate that our model significantly outperforms state-of-the-art methods.

With the explosive popularity of mobile devices, mobile crowdsensing (MCS) has emerged as a promising large-scale data collection paradigm. Suitable incentive mechanisms are essential for encouraging user participation. Current MCS work more or less ignores three factors. First, mobile users are assumed to be independent of each other, ignoring social effects. Second, due to the heterogeneity of users, if you just blindly attract users without distinguishing them, the data quality can decrease. Finally, the limited communication resource allocation problem during uploading sensing results is ignored. Therefore, we model the quality-improved and delay-aware incentive mechanism with social concerns as a two-stage Stackelberg game, in which the rational use of social effects not only motivates user participation but also avoids a serious decline in information value due to repetition, and reasonable allocation of communication resources ensures the timeliness of delay-sensitive tasks. Furthermore, data screening, similarity analysis, voting, and reputation are used simultaneously to improve data quality. The Hessian matrix in a multiuser, multitask hyperspace setting is utilized to verify the existence and uniqueness of the game equilibrium. The closed-form expressions of the optimal requester pricing and the optimal user data load strategies are derived, respectively. The proposed mechanism is compared with gather–scatter, incentive-G, Blockchain-based secure, interactive, and fair MCS (BSIF), and Socially-aware incentive mechanism (SAIM) algorithms. Extensive simulation results on a real trajectory dataset show that compared with these state-of-the-art algorithms, the proposed incentive mechanism can motivate users to provide more data loads with few rewards, greatly improve the requester utility, and suppress the data upload of malicious users.

Mobile crowdsensing (MCS) has recently shown good performance in solving large-scale sensing tasks. As an essential topic in MCS, recommending tasks to participants has received extensive attention from researchers. Most studies assume that participants are absolutely rational, which is unrealistic because it is difficult for participants to know all the information about the transaction. Furthermore, most of them do not consider how to learn the preferences of new participants. In addition, their works are difficult to extend to different MCS scenarios. Considering the above problems, we propose an extensible bounded rationality-based task recommendation scheme (EBRTR), which contains a task recommendation framework and a bounded rationality decision-making model. First, a task recommendation framework that can be easily extended to various MCS scenarios is designed. Second, in our bounded rationality decision-making model, for participants with historical task information, according to the implicit information in their historical tasks, the human thinking mode with bounded rationality is simulated, and the improved classification and regression tree (ICART) algorithm is designed to construct the decision tree. For participants who newly join the platform, social information is introduced to construct an initial decision tree. Finally, extensive experimental evaluations demonstrate the effectiveness of the proposed scheme.

Image segmentation is an essential part of medical image processing, which plays a significant role in adjunctive therapy, disease diagnosis, and medical assessment. To solve the problem of insufficient extracting context information, especially for medical image segmentation, this paper proposes a novel network architecture of multi-scale object context dilated transformer network (Multi-OCDTNet) to improve the utilization and segmentation accuracy for context information. The multi-scale object context transformer module can extract the multi-scale context information of the image through a three-layer transformer structure in a parallel way. The dilated convolution self-aware module can enhance the awareness of multi-scale context information in the feature map through layering transformer block groups and a set of transformer layers. In addition, we propose a composite weight-assigned-based loss function based on DDCLoss and Focal Tversky Loss to improve the stability of the segmentation performance of Multi-OCDTNet by adjusting the weight. The performance of Multi-OCDTNet is validated on the DRIVE and STARE datasets with segmentation accuracy of 97.17% and 97.84%, respectively, indicating the Multi-OCDTNet network possesses a significant competitive advantage in improving the segmentation performance of retinal vessel images.

With the rapid development of smart cities, vehicles equipped with various sensors can effectively sense traffic, thus forming a crowdsensing paradigm for the intelligent transportation system (ITS). Although mobile crowdsensing in ITS has broad application advantages, it still faces many challenges, such as single point of failure, inefficient independent task allocation, and the inability to deal with safety emergency tasks in time. To handle the abovementioned issues, we establish a decentralized ITS architecture based on blockchain and propose the concurrent tasks assignment problem proved to be NP-hard and safety emergency tasks assignment problem. Then, we propose reinforcement learning-based concurrent tasks and the safety emergency tasks assignment method, which can maximize the utility of concurrent tasks based on satisfying the requirements of safety emergency tasks. Simulation results demonstrate the effectiveness of the proposed methods.

Web technology underpins many interactive mobile applications. However, energy-efficient mobile web interactions is an outstanding challenge. Given the increasing diversity and complexity of mobile hardware, any practical optimization scheme must work for a wide range of users, mobile platforms and web workloads. This paper presents CAMEL, a novel energy optimization system for mobile web interactions. CAMEL leverages machine learning techniques to develop a smart, adaptive scheme to judiciously trade performance for reduced power consumption. Unlike prior work, CAMEL directly models how a given web content affects the user expectation and uses this to guide energy optimization. It goes further by employing transfer learning and conformal predictions to tune a previously learned model in the end-user environment and improve it over time. We apply CAMEL to Chromium and evaluate it on four distinct mobile systems involving 1,000 testing webpages and 30 users. Compared to four state-of-the-art web-event optimizers, CAMEL delivers 22% more energy savings, but with 49% fewer violations on the quality of user experience, and exhibits orders of magnitudes less overhead when targeting a new computing environment.

Currently, as one of the most important video coding structures, the low-delay (LD) configuration has been widely adopted in the video communication systems, such as video surveillance and video conferencing systems. The rate-distortion optimization (RDO) is also a crucial technique for improving the performance of video coding. The RDO performance of H.265/HEVC with the LD configuration is significantly affected by the temporal dependence. This paper examines the dependence of rate and distortion among temporal layers in H.265/HEVC with the LD configuration. An efficient RDO method is then proposed based on temporal dependence, where the optimized Lagrangian multipliers are estimated by adaptively selecting the scaling factors according to the video motion characteristics pertaining to different temporal layers. The experimental results show that compared with the RDO method adopted in the H.265/HEVC test model (i.e., HM), the proposed RDO method can achieve the average BD-rate reductions of 6.4% and 5.8% for the LD P and LD B configurations, respectively.

For adaptive region-based despeckling methods, the shape of the adaptive region is of large influence on the estimation of the true value. In this paper, given the effectiveness of patch-based index on measuring the relativity of samples, an adaptive region is formed at each pixel via the patch-based index where the local geometric correlation (such as the anisotropy and the compactness) among the pixels in the patch is explored as an Adaptive Kernel(AK) function. Then, using all pixels contained in the adaptive region, Maximum likelihood rule is adopted to estimate the true value of the concerned pixel. From the experimental results on real and synthetic SAR images, by using the AK function embedded patchy index to determine the adaptive region, not only the speckle is largely reduced, but also the resolution of the details is well preserved by our method.

The influence of noises is obviously in the image transform method based on the pixels elements representation of images. To some extent the method of image transform based on linear elements representation can solution the problem. The Beamlet transform is an effective method of line segment extraction. This work improved the traditional Beamlet transform by considering the directional information of lines. The improved method transformed a digital image to a coefficient matrix. This method can embody the linear singularity of some linear targets and can be used for edge detection and extracting other useful targets in noisy images. The experimental results on manual images and SAR images demonstrate the effectiveness of this method.

In this paper, we propose a cross-layer optimization framework to jointly design the spectrum sharing and flow routing with the interference considerations in cognitive radio networks. Given multiple traffic demands from different source nodes to destination nodes, we formulate an optimization problem in the form of mixed integer linear programming (MILP) to provide a fair routing. Different from the existing work, we consider bi-directional links because we believe the link level acknowledgements in an ad-hoc network are a must. For traffic routing we allow multi-path for each traffic demand. Numerical results show that the spectrum sharing among the secondary users is interference-free and a fair routing is guaranteed for different traffic demands.

Cognitive radio technology solves the spectrum under-utilization problem by enabling the secondary users access the spectrum holes opportunistically. Therefore, how to efficiently share the spectrum holes among the secondary users is of interest. Previous studies on spectrum sharing focused on the formulations with homogeneous channels. The channel heterogeneity, which is a unique feature in cognitive radio networks, has been ignored. In this paper, we consider heterogeneous channels and explore the impact of channel heterogeneity on the spectrum sharing. We model the channel heterogeneity, interference constraints and spectrum sharing, and formulate an optimization problem in the form of binary integer linear programming (BILP). To the best of our knowledge this is the first attempt to model the channel heterogeneity into the formulation of spectrum sharing in cognitive radio networks. Numerical results show that the optimal solution on spectrum sharing is highly dependent on the channel heterogeneity.

A major threat in wireless sensor network is report fabrication attack. The existing approaches in which each forwarding sensor along a path probabilistically filters out injected false data may not be adequate because such protection may break down when more than a threshold number of nodes are compromised. As an alternative, we present a sink filtering scheme in clusters of heterogeneous sensor networks to solve the threshold breakdown problem. Besides basic sensors, we add some powerful data gathering sensors, termed as cluster heads (CHs). To be accepted by the sink, each aggregation report generated by a CH must carry multiple keyed message authentication codes (MACs); each MAC is generated by a basic sensor that senses the event. We analyze the resilience and overhead of the sink filtering scheme. Both analytical and simulation results show that the scheme is resilient to an increasing number of compromised nodes, with graceful performance degradation. The scheme is also scalable and efficient in communication, computation and storage

In this paper, we propose an IP onetime-password authentication scheme over insecure data channel, for resource access control at an edge router. We assume a master secret key is pre-established between a legitimate end-host and an edge router. Prior to data transfer, a four-message-handshake scheme for setup is applied over insecure data channel, through which a legitimate end-host communicates the commitment of a one-way key chain (i.e., the rst key) with the ISP edge router. The legitimate end-host then uses the keys of the key chain as the onetime-passwords in the order reverse to the generation of these keys. The edge router, knowing the correct sequence of the valid onetime-passwords, veri es the received onetime-passwords and denies undesirable traf c to access the reserved network. The authentication scheme of IP onetime-password is secure even if an attacker can temper with and eavesdrop on the communication between end-host and edge router.

As the Internet becomes pervasive, the vulnerability of some fundamental design aspects of the Internet has also become significant. Among which, denial-of-service (DoS) and distributed DoS (DDoS) pose significant problems, as they are disruptive to the useful traffics and are hard to prevent. One solution consists in instituting accountability, which hold the attackers accountable for the attack. The key issue is to identify the real sources of the attacks and attackers use spoofed IP address to hide their actual network location. However, the Internet architecture does not provide intrinsic support for identifying the real sources of IP packets. Numerous mechanisms have been proposed to traceback the real sources. Most of such networks have been addressing the IP version 4. In this paper, we address the issues of IP traceback in the context of IPv6 and mobile IPv6. This paper provides a detailed analysis of these issues and problems. The main problem lies with the transformations that are introduced by IPv6 and mobile IPv6 protocols, namely tunneling and addresses manipulation. We then propose a solution, including new ICMPv6 messages for traceback co-ordination, to facilitate the traceback mechanism.

In this paper, we study a fixed wireless access (FWA) system supporting both fixed and mobile users. A multi-dimensional Markov chain is developed which captures the mobility and call arrival pattern of the users. Three different policies including complete sharing, guard channel and reserved trunk are evaluated. Analytical results are provided to demonstrate the effects of various parameters on the system performance.