Federated learning (FL) has been increasingly applied to smart city applications to enable distributed model training while preserving data privacy. However, conventional weight-sharing FL still incurs high communication costs in bandwidth-constrained edge environments. Recently introduced Federated Prompt Learning (FPL) significantly reduces communication bandwidth usage to alleviate this issue. Nevertheless, performance degradation persists due to the coexistence of inference and training tasks and communication overhead, defined as delays arising from the transfer of tasks and associated data. Managing priorities between latency-sensitive inference and delay-tolerant training tasks is a critical challenge.This paper models the offloading of inference and training tasks in FPL environments using queuing theory. Compared to the Task-Type Separation (TTS) method, we expand the control model and propose an improved offloading scheme that minimizes the mean sojourn time of training tasks while meeting delay requirements for inference tasks. This scheme allows for bidirectional offloading among multiple edge servers, reducing inference latency while maintaining training continuity. Analytical evaluations show that the proposed method satisfies inference delay requirements while maintaining acceptable training delays. This provides an automatic, scalable task control mechanism for FPL-based edge systems.

In recent years, the low earth orbit (LEO) satellite network has attracted much attention. LEO satellites enable low latency communications, however the communication range of individual satellites is narrow, requiring a constellation of many satellites. The limited number of satellites in the initial phase of service provisioning limits the service area and communication time, making it difficult to secure users. In this paper, we propose a model in which multiple LEO satellite operators cooperate to provide services, and the revenue is reasonably allocated to each service provider based on the Shapley value of cooperative game theory. The possibility of extending communication services through cooperation among operators is verified through numerical evaluations.

Mobile Edge Computing (MEC) is attracting attention as a solution to increasing network traffic load. To build an MEC system, edge servers are placed on base stations and the users must be allocated to edge servers. The conventional MEC method places edge servers and allocate user jobs by using extension of k-means to minimize the number of hops between a user and servers. With this method, however, heterogeneous environments are not assumed, which means the performance difference among edge servers. When many jobs arrive at edge servers with low performance, jobs overflow and transfer to the cloud servers. The initial value dependence of this conventional method has also not discussed. We propose a method for initial edge-server placement and job reallocation in order to apply heterogeneous edge servers and initial value dependence.

Image watermarking is an important tool for copyright management, and various watermarking methods have been proposed. Tensor decomposition, which has attracted much attention in image processing, has been applied as a watermarking method, however, no method has been proposed for color cover images with grayscale watermarks. Therefore, we propose a color image watermarking method that involves using a grayscale watermark. We evaluated the attack resistance of the proposed method through experiments.

The recent surge in traffic demand has accelerated the development of heterogeneous high-density networks. Consequently, the number of base stations has increased, leading to a rise in the frequency of vertical handovers between heterogeneous base stations, which results in delays and increased power wastage. Earlier studies analyzed the number of vertical handovers in such networks. Unfortunately, while the number of vertical handovers with short sojourn periods is increasing, the sojourn time itself has not been well addressed. This paper prepare to analyze vertical handover with direct consideration of the short cell sojourn time distribution as derived from stochastic geometry. Additionally, by referencing Bertrand’s paradox, we define and simulate randomness in three different scenarios to validate our findings.

Video distribution services have become widespread and are used by many people. Therefore, the shared link becomes a bottleneck, and the quality of experience (QoE) deteriorates. Therefore, DASH (Dynamic Adaptive Streaming over HTTP) is used to maintain QoE by dynamically selecting image quality. A conventional method improves QoE by dynamically selecting the playback time (segment length) of one segment by calculating the variation of client throughput in addition to the selection of image quality. However, the segment at the beginning of playback did not calculate the throughput variation properly, and the QoE decreased due Åto inappropriate segment length selections. In this paper, we propose a method to improve the QoE by selecting the segment length at the beginning of playback using the bandwidth of the bottleneck link and the number of connected users, which can be measured in advance, and by changing the derivation method of the throughput variation.

For multi-transmission rate environments, access point (AP) connection methods have been proposed for maximizing system throughput, which is the throughput of an entire system, on the basis of the cooperative behavior of users. These methods derive optimal positions for the cooperative behavior of users, which means that new users move to improve the system throughput when connecting to an AP. Typically, the arrival of two or more new users causes signal interference for other users. Specifically, traditional methods have not considered interference and the number of allowable mobile positions is so large that it is difficult to find the optimal position to maximize system throughput. In this paper, we propose an AP connection method that maximizes system throughput by considering interference between users, with a potential game model. This game can find at least one optimal user position. Moreover, numerical analyses show the effectiveness of our method.

Perceptual hashing, which generates a message digest showing how humans perceive similarity in images, is suitable for ensuring the equivalence between a modified/edited image and the original. Conventional perceptual hashing is mainly utilized for similarity-based image retrieval and is not appropriate for image identification, which is required in security systems such as digital rights management. We previously developed a construction method for perceptual hashing in security systems that utilizes a convolutional neural network (CNN). In practical applications, multiple different images are published in various media forms (e.g., articles or books), so generating an identical message digest for each of these images simultaneously makes content management easier. Therefore, in this work we extend our earlier CNN-based perceptual hashing scheme so that it can generate an identical message digest for images in a group. This approach reduces the computational cost for fine-tuning CNN compared to generating a perceptual hash for each image in a group individually.

Abstract Many network devices, such as a smartphones and tablets, and streaming services, such as YouTube and Netflix, have become widespread. This causes network‐bandwidth congestion, so users require stable and consistent communication quality. Call admission control (CAC) has been proposed to solve these bandwidth resource problems. Some CACs have been proposed in order to maintain QoS. However, these CAC does not assume user behaviour. If CAC assumes some user behaviour, optimal control parameter derived conventional CAC may change because traffic load change by considering user behaviour. This article assumes that arriving types of flows are of three types: narrowband flows, broadband flow without seek‐bar operation, and broadband flows with seek‐bar operation. Under this assumpution, two CAC methods with single threshold or two thresholds are proposed. With these methods, it is assumed that all arriving flows result in the same user satisfaction when accommodated in a network. Under this assumption, the maximum of each user satisfaction is the same as the maximum number of accommodated flows. Using the call loss probability characteristics of Method 1, the objective function in Method 2 is modified. By using the modified objective function, CAC can be realised while properly accommodating broadband users who have been excessively rejected.

In a P2P MMORPG, the game space is divided into multiple areas, and nodes are placed in each area to manage the game. If the load exceeds the processing capacity of the nodes, the game will be interrupted. Thus, the load needs to be distributed. A conventional study for a load balancing algorithm that detects the movement of players has been proposed to solve this problem. This method dynamically divides and combines the sub-areas by setting two thresholds: one for the upper limit and the other for the lower limit of the number of players that can exist in each sub-area. However, the load increases as the number of divided areas increases. In this study, we propose a novel load balancing algorithm to reduce the load using k-means++.

Perceptual hashing generates a message digest based on the image content of the human visual system, and differs from the cryptographic hash function that generates the hash value based on each bit of the image file. We apply perceptual hashing to digital watermarking to generate watermark information after each image modification/editing, and verify that modified/edited images and the original image are the same in copyright. To obtain a stable perceptual hash value robust to image modification/editing for digital watermarking, we previously developed a construction method for perceptual hashing using a convolutional neural network (CNN). This was necessary because the conventional perceptual hash algorithms are used for database retrieval, and the required characteristics are different from those used for digital watermarking. However, in this method we needed to fine-tune the CNN for each image used to calculate the perceptual hash value, which led to inefficiency. In order to make the calculation of the perceptual hash value more efficient, we propose a construction method for perceptual hashing based on CNN that does not require fine-tuning. In the proposed method, an image is input to the CNN and the perceptual hash value is calculated based on the response of the output layer of the trained CNN.

We propose a flow admission control method to maximize the satisfaction of all users when accommodating communications with different bandwidths in the same network, and we demonstrate its effectiveness using theoretical equations. Conventional flow admission control assumes the rationality of all users requesting the correct bandwidth. However, in a real network, some users may mistakenly select the incorrect bandwidth. Thus, we need to analyze user behavior to calculate an optimal control parameter for flow admission control. We model our flow admission control with user behavior by using stochastic evolutionary game theory and queueing theory. In addition, we analyze the characteristics of our flow admission control while accounting for user behavior. As a result, we find the total user satisfaction is maximized when the selection error probability is 20% under the conditions of our study.

Digital watermarking technology is used extensively in the field of digital rights management. However, there are a few problems when it comes to making effective use of digital watermarking. First, for conventional digital watermarking, a digital image is used only as a carrier for embedded watermarking information, and as this information may be diverted to other images, the watermark information needs to be generated based on the original image. Second, after the original image is modified/edited, the watermark information needs to prove that it is from the original image. Third, multiple digital watermarks need to be stored and managed without depending on trusted third parties. In an earlier work, we proposed a digital rights management system based on digital watermarking, blockchain, and perceptual hashing to resolve these issues. However, because we used conventional perceptual hashing, we could not draw sufficient conclusions about the first and second problems. In order to obtain a stable digest message of an image for digital watermarking, we here propose a new construction method for perceptual hashing using a convolutional neural network (CNN). In the proposed method, we first construct a machine-learned CNN for accepting an image that we want to take the perceptual hash value. The perceptual hash value is the cryptographic hash value of the weights that make up the CNN. We then verify that the reconstructed CNN can guarantee the hash value used when obtaining the hash value, and confirm that the image to be verified is accepted and is the perceptual hash value of this image.

In recent years, the available network bandwidth is decreased by increasing the mobile devices such as a smartphones or tablets. Quality of Service (QoS) control is required to guarantee the communication quality for users of the network. As one of QoS control techniques, which judges a new arrival streaming application (=flow) can be accommodate in the network has been proposed. Conventional admission control methods that focus on the user's cooperative behaviour have been proposed. In general, some users use the video navigation tools like a seek-bar to jump to different time position because they want to watch some specific scenes. However, this conventional method does not assume the user's behaviour. In this paper, we propose an admission control to maximize user satisfaction by considering user behaviour to reduce the video time. We evaluate our proposed method by numerical analysis using queuing theory and show the effectiveness of our proposed method.

In the past, the improvement of digital copyright protection system based on digital watermarking mainly focused on algorithms, while generation and storage of the watermark information was ignored. In this paper, a new design scheme of copyright management system based on digital watermarking and its information, such as blockchain, is proposed, which combines digital watermarking, blockchain, perceptual hash function, Quick Response(QR) code, and InterPlanetary File System(IPFS). Among them, blockchain is used to securely store watermark information and provide timestamp authentication for multiple watermarks (multiple copyrights) to confirm the creation order. Perceptual hash function is used to generate hash value based on the structure information of images, that watermark information can be confirmed without the original image. QR code is used to generate QR code images containing image hash and copyright information as watermark images to improve robustness and capacity of digital watermarking; IPFS is used to store and distribute watermarked images without a centralized server. This scheme can enhance the effectiveness of digital watermarking technology in the field of copyright protection. In this way, use P2P network to integrate and complete copyright management and distribution of copyrighted works without requiring a trusted third party. Nodes rely on cryptography to confirm the identity of each other and ensure the security of information. It can reduce information leakage, data destruction and other risks caused by collapse of the centralized system in the past. This improves the security and transparency of information, and speeds up the distribution of copyrighted works to facilitate circulation in the network. This scheme can also improve copyright protection of multiple creations. Combine blockchain and multiple digital watermarks to record copyright information of every copyright owner in the authoring process and fully prove this information. In order to protect the legitimate rights and interests of each copyright owner.

The long reach passive optical network (LR-PON) is a promising scheme for access networks that cover large areas. For designing an access network, an exact solution of the network performance (e.g., mean packet delay) is useful. Since long propagation delay causes unused time slots (idle time), and its effect on performance degradation is difficult to analyze, existing analysis of the LR-PON is done only for the no-idle time situation with modified optical line terminal (OLT) and each optical network unit (ONU). In this paper, we propose an exact solution of the mean packet delay for LR-PON with idle time. Our analytical solution is derived by a non-trivial extension of the existing work for the short range PON. We confirm a good match between the mean packet delay derived by the simulation and our analytical solution.

We propose an optimal access-point (AP) selection algorithm for maximizing system throughput while preserving newly arrived-user throughput. In our algorithm, newly arrived users cooperate with a WLAN system they are trying to use, i.e., they are willing to move toward an appropriate AP. We discuss the effectiveness of our proposed AP selection algorithm by using numerical analysis. We also clarify and analyze the characteristics of system throughput.

We propose a novel flow admission control that minimizes the total call-blocking rate by treating heterogeneous flows equally. We focused on the static-state bandwidth load, which refers to how much bandwidth is used statically by a smaller class of flow than a newly arriving flow. In this paper, we analyze the characteristics of different types of loads, explain the development of a call-blocking curve to minimize the total call-blocking rate, and demonstrate good performance by this call-blocking curve.

We have developed a novel call admission control (CAC) to minimize the total call-blocking rate on the basis of a new philosophy in which heterogeneous traffic is treated equally in network. We also obtained optimal threshold that minimizes the total call-blocking rate. However, to obtain this optimal threshold, we needed to full-search all threshold candidates. In this paper, we propose an approximation formula to obtain suboptimal threshold based on theoretical characteristic analysis and demonstrate that this approximation formula performs well.