With the great development in smart devices, mobile crowdsensing (MCS) has been an innovative paradigm for data gathering. Task assignment is a fundamental problem in MCS systems and applications. Previous studies only focused on the assignment of individual tasks, neglecting planning the task processing from a higher level, e.g., making assignments between task locations and workers, which impacts the crowdsensing performance adversely. Furthermore, task characteristics, e.g., route distance, task similarity, and task priority, have a great impact on the rationality of a visiting order and the quality of crowdsensing services. In this article, we tackle the problem of rational task assignment and path planning for MCS, which aims to assign a set of task locations to a set of workers and generate location visiting sequences. We measure the assignment rationality by taking into account geographical information and task characteristics, i.e., route distance, task similarity, and task priority. We prove that the problem of computing the rationality maximization task assignment is NP-hard. For the single-worker scenario, we reduce the problem to a simpler problem with respect to only route distance and task priority criteria since the similarity measurement has a fixed value. We propose an effective greedy algorithm. For the multiple-worker scenario, we first extend the greedy idea by considering all three criteria and then propose an effective approach by reducing the computational complexity of similarity measurement. Extensive experiments show that our proposed approaches achieve promising results.

Fog computing is used to process a large amount of data produced by terminal devices where fog nodes are the closest acquirers to the terminal devices. In fog computing, the processed data may be tampered with or illegally captured by some malicious nodes while the data is transferred or stored. So, when some applications need to require data process with high security, fog computing must provide a security approach to secure and check the final results. In this paper, we propose a secure data sequence query framework based on multiple fog servers, where we use multiple fog servers to store and provide data hashing values to verify corresponding data sequence. In the proposed scheme, the cloud server needs to pre-designate some Merkle hashing tree topologies to the fog network, then the fog server directly acquires related data from leaf-nodes (fog nodes) according to one of the pre-designated Merkle hashing tree topologies; at the same time, an actual Merkle hashing tree is constructed according to the selected Merkle hashing tree topology in the fog network. In fog computing, the related fog nodes participate in computing and transferring the hashing values on data, then the corresponding root node uses the $(t,n)$ threshold secret sharing scheme to compute and distribute the sub-secrets of the root value to other fog servers. Therefore, when a user requests a data sequence from a fog server, the fog server may immediately respond to the request and the other fog servers can provide the sub-secrets of hashing root value to verify the corresponding data sequence. Our proposed framework can ensure that users may obtain required data fast, accurately and reliably. Additionally, based on the data query security requirements of fog computing, we analyze the security of our proposed scheme. Our proposed scheme not only guarantees the reliability of data but also effectively protects data against various attacks.

Vehicular networks, which are recognized as an innovative technology for information collection due to the powerful sensing capability and strong mobility, have been an important platform for geographic crowdsourcing services and are particularly useful in environmental data collection. Crowdsourced data collection in vehicular networks must be performed in a communication-efficient manner due to the resource-constrained wireless communication links. Moreover, it is a pressing problem to improve the quality of responses because workers may provide data of poor quality or even fabricate data to defraud rewards. While the existing work has focused on spatial/temporal task coverage and monetary rewards, in this work, we propose a cost-efficient framework for crowdsourced data collection in vehicular networks while ensuring the accuracy of the response. Crowdsourced data collection consists of two main steps: 1) task assignment and 2) crowdsourced answer gathering. We first propose a task assignment scheme that maximizes the overall data quality and reduces the amount of data transmission. We estimate the data quality level based on a Gaussian mixture model and reduce the amount of data transmission by carefully selecting a subset of vehicles for crowdsourced tasks. We then design an answer gathering scheme that considers both the length of the aggregation tree and data delivery and minimizes the communication cost for collecting answers from participants. Extensive experiments on both synthetic data sets and real data sets show that our proposed framework achieves promising results.

Individuals can easily generate highly realistic images using artificial intelligence-generated content technologies, which complicates the verification of images’ ownership rights. This raises potential issues such as spreading misinformation, fraud, and copyright infringement. Digital watermarking is a promising solution to protect the copyright of a digital image by embedding watermarks within it. However, many existing deep learning-based watermarking approaches struggle to simultaneously resist multiple attacks effectively and maintain the quality of watermark images. In this paper, we propose a bidirectional-interactive and context-aware (BICA) deep network designed to enhance the robustness of the watermark while maintaining the quality of the encoded image. We propose a new attention module in the encoder to improve the invisibility and robustness of the watermarked images by implementing an adaptive two-way interaction between local and global features. Additionally, we employ fine-grained downsampling to enhance the attention module’s ability to capture comprehensive feature information. Extensive experimental results demonstrate that the BICA network can embed watermark information into an image without compromising image quality. For instance, BICA has an accuracy exceeding 95% against various moderate noise attacks, with average PSNR and SSIM values of 40.4021 dB and 0.9943, respectively.