Non-terrestrial networks (NTNs) contribute significantly to offering connectivity services for the upcoming sixth generation (6G). However, many services demand real-time uplink data transmission. Owing to the dynamic topology, instability of NTN channel status, resource constraints, and tremendous mobile users on ground networks; resilient and effective resource management and channel accessibility pose substantial challenges. This work investigates the uplink sum-rate maximization and latency minimization problems with a channel allocation scheme for user equipments (UEs) to high-altitude platforms (HAPs) of a particular NTN when UEs simultaneously transmit data to HAPs and HAPs forward it to low-earth orbit satellite (LEO-sat) via multiple subcarriers. We formulate a fairness resource allocation optimization problem reflecting the system utility and load of the HAP-LEO-sat link for aerial-to-space uplinks. To tackle the high-dimensional non-convex optimization problems and short visibility of LEO-sat in the learning, we advocate a novel deep reward-based multi-agent reinforcement learning (DeepRMARL) with prioritized experience replay (PER) and priority sampling approach (PSA). Furthermore, we present a long-short-term memory (LSTM) model for future environment channel prediction due to time-varying channels between UEs and HAPs. The results explicitly unveil the feasibility of our method in optimizing total uplink rate, achieving a higher fairness index, minimizing latency by approx. 12.2%, and predicting future environment states with approx. 92.8% accuracy.

Non-terrestrial networks (NTNs) are crucial for 6G communications but face challenges like classical communication bottlenecks. Semantic communication (SemCom) offers a solution, but efficient resource allocation remains unresolved. This paper proposes a joint approach for resource allocation and semantic information selection in low-earth orbit (LEO) satellite-based image transmission. A novel metric, Quality of Semantic (QoSem), is introduced to measure the SemCom performance. An optimization problem is formulated to maximize QoSem by allocating computing, power, and bandwidth resources and selecting optimal semantic information. Furthermore, we propose a novel deep reinforcement learning (DRL) algorithm, namely Mixed-Action Decision Intelligent Semantic-aware Soft-Actor Critic (MADIS-SAC), to handle the problem’s nonconvexity and NP-hard complexity. Results show 82.5% of QoSem improvement compared to baselines under resource constraints.

Abstract This paper addresses the issue in live events, especially in sports viewing, where individuals who are deaf, hard of hearing (DHH), blind, or have low vision (BLV) struggle to access sufficient information. Traditionally, information accessibility has relied on specific professionals or volunteers, which is often inadequate. To tackle this challenge, we propose a mechanism that facilitates information sharing for all individuals, regardless of their abilities. We have developed an inclusive web application tailored to address these needs. This application is beneficial not only for DHH and BLV individuals but also for all audiences. Additionally, the system’s applicability extends to other domains, such as museum visits. This paper details the newly developed web application and presents the outcomes of pilot studies conducted in sports viewing and museum settings with DHH and BLV participants. The results of these experiments are analyzed to assess the system’s effectiveness and to identify future improvement areas.

Multicast-based multimedia streaming is expected to be a key application in the next generation of wireless and mobile networks. For such applications, we utilize hierarchical non-terrestrial networks (NTNs) consisting of low earth orbit (LEO) satellites, high-altitude platforms (HAPs), and user equipment (UE). However, existing multicast protocols for NTN cause a long delay and jitter owing to packet loss, thus being the cause of low-quality streaming services. We propose a reliable multi-cast protocol, namely, overhearing and non-orthogonal multiple access (NOMA) repairing reliable multicast (ONRM), for delay reduction. In ONRM, the HAP overhears packets from the LEO to UE and retransmits the lost packets using NOMA on behalf of the LEO to prevent delay increments. We carried out evaluations based on the release of Third Generation Partnership Project (3GPP) in terms of delay, throughput, and jitter. We prepared existing multicast protocols, pragmatic general multi-cast (PGM), reliable and efficient multicast protocol (REMP), and extended bit-indexed explicit replication (extended BIER) for comparison. Results show that ONRM outperforms other multicast protocols. Especially, the proposed ONRM can realize a low jitter and enables high-quality multimedia streaming over NTN.

Inter-cell interference coordination (ICIC) has been widely studied for mitigating the effects of severe inter-cell interference (ICI) in cell- edge users. However, based on the scarcity of frequency resources in orthogonal multiple access systems, the ICIC methods proposed in the previous papers have difficulty in maintaining the overall performance and fairness of a system. Non-orthogonal multiple access (NOMA) is a promising radio access technology that can serve multiple users simultaneously with the same frequency resources. However, most previous work has not considered the ICI problem in NOMA systems. We propose a centralized adaptive ICIC framework for downlink NOMA systems, including a distributed clustering algorithm, a distributed power allocation algorithm, and a centralized frequency allocation algorithm. Simulation results demonstrate that the proposed framework outperforms all benchmark frameworks and can improve both the overall performance of a system and fairness among users.

In 360-degree video streaming, a user may watch a part of video frames, namely, viewport, through an interactive display, e.g., head-mounted display, for the immersive experience. One of major issues for high-quality 360-degree video delivery is how to reduce perceptual redundancy in video frames according to user's viewing viewport. To this end, an existing scheme divides video frames into multiple tiles and adaptively encodes each tile based on user's viewing viewport using digital-based video compression. However, the video compression causes an issue in wireless channel called cliff effect, which causes sudden quality degradation in user's viewport and poor immersion in immersive applications. To reduce perceptual redundancy without cliff effect in wireless 360-degree video delivery, we propose a novel transmission scheme. Our scheme skips quantization and entropy coding, instead, directly transmits linear-transformed signals based on three-dimensional discrete cosine transform (3D-DeT) or the combination of one-dimensional DeT (1D-DeT) and spherical wavelet transform (SWT). By skipping both operations, the quality of viewport can be simply enhanced by its power allocation since it is directly applied at the pixel level instead of the bit level. It is demonstrated with 360-degree video sequences that the proposed scheme offers a great advantage over the conventional digital-based schemes.

A project to develop a blind football play-by-play system was started. The target of the system is visually impaired spectator. To conduct a basic study, we made an experimental system that detects a position of a player using a Quasi-Zenith Satellites System (QZSS) and shows the position via refreshable tactile display. The QZSS allows users to acquire more precise position information than only with GPS around Japan. The study showed that a certain possibility to acquire an arbitrary position data of the player in the field and display it to the tactile display.

We have promoted a project called "Information Support of Everyone, by Everyone, for Everyone (ISeee). "We have been developing a live timeline posting/reading system to use at sports games for audiences to share data consisting of photos and texts. We describe two experiments using the system for the purpose of information support for deaf and hard-of-hearing people (DHH). The results suggest that the data generated with the live timeline system is helpful as information support for DHH.

With the increasing popularity of smartphones and Internet of Things (IoT) devices, there has been an increase in the amount of mobile traffic on networks. In this paper, we discuss an access strategy for downlink/uplink decoupling (DUDe) with multiple frequency channels to enable the efficient use of frequencies in wireless networks. DUDe is a wireless communications scheme that effectively utilizes frequencies when communicating by connecting the uplink and downlink to different base stations. We propose two base station selection methods, i.e., signal-to-interference-plus-noise ratio (SINR)-based decoupling with re-association (SBD-RA) and SINR-based decoupling with first-come first-association (SBD-FCFA). Based on the evaluation results obtained by event driven simulations, we show that both SBD-FCFA and SBD-RA provide higher throughput by utilizing frequencies more efficiently than DUDe when using the same channel, or without using DUDe. We also show that SBD-RA achieves a high throughput although SBD-RA has a larger computational complexity than SBD-FCFA.

Current one-to-one wireless communications is reaching the Shannon limit. Previous works have studied space multiplexing schemes, e.g., superposition coding and successive interference cancellation, to overcome the limitation. In this paper, we focus on a network multiple-input multiple- output (MIMO) architecture, which is one such space multiplexing scheme, where multiple access points in the network MIMO system cooperate with each other to improve wireless communications capacity. However, channel sounding overhead for estimating the channel state information of every path between multiple access points and multiple clients is a significant problem. Likewise, the computational overhead for deciding which paths to use for the network MIMO transmission is also high. We propose Throughput-Aware Random Clustering (TARC) of access points to reduce the network MIMO overhead. TARC takes a cross-layer approach for choosing access points to participate in network MIMO transmission on the physical layer based on throughput in the data link layer. From our validation using simulations, we show that the proposed method is able to achieve approximately 2.5 times higher throughput than using all access points in the network and 1.4 times better throughput than adopting static cluster sizes.

There are many pipes buried underground in urban areas. An installation of signal posts requires information about actual underground conditions in order not to crack any pipes. However, some areas do not have the accurate pipe layouts because the constructions make a gap between the layouts and the results. Using a ground-penetrating radar (GPR) is a solution for preventing damage to the pipes regardless of the accuracy of layouts. Our goal is to propose an automatic pipes-detection method using the GPR signals. The achievement provides the following two things: streamlining a survey of the underground without expert's experience and reducing a burden on the users. In this paper, we propose a new detection method based on the Dirichlet process mixture (DPM) model. This paper aims at examining the estimation accuracy of our method. First of all, the method of our previous work reduces noises of the GPR signals. Secondly, the two-dimensional Gabor wavelet (2D-GWT) is applied for the denoised signals. Next, samples are drawn from the 2D-GWT result which we regard as a probability distribution. Finally, we obtain the partition of samples by using the fixed DPM model to be proposed. We call it the Dirichlet Process Crescent-signal Mixture model. We estimate the positions of buried pipes from the partitions. Some estimated positions are close to the true ones. However, the estimated depths tend to be greater than the true ones because the relative permittivity of underground is apt to increase. We find that the constant relative permittivity is an erroneous assumption. This issue for precise estimation will be addressed in our future research.

The purpose of our research is to improve the recognition rate of offline character recognition systems using the HMM (Hidden Markov Model) without increasing the number of HMM parameters too much. Some 2-dimensional HMM character recognition systems have been proposed to increase representational power. However since 2D HMM has a much more complex structure and thus requires much more parameters than 1-dimensional HMM, it becomes very hard to gather sufficient samples in order to guarantee successful generalization. To overcome the problem, we propose a method for character recognition using 1D HMMs in multiple directions with 2-dimensional feature extraction. To further improve the performance, some voting methods using a bagging algorithm are also exploited. In our experiment, the recognition rate is increased by about 1% with the multiple directional HMM character recognition system compared to the 1D HMM character recognition system. The recognition rate is further increased by about 1% with the HMM character recognition system using a bagging algorithm.

Article Collaborative customer services using synchronous Web browser sharing Share on Authors: Makoto Kobayashi IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, Japan IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, JapanView Profile , Masahide Shinozaki IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, Japan IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, JapanView Profile , Takashi Sakairi IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, Japan IBM Research, Tokyo Research Laboratory, 1623-14, Shimotsuruma, Yamato, Kanagawa 242-8502, JapanView Profile , Maroun Touma IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NY IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NYView Profile , Shahrokh Daijavad IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NY IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NYView Profile , Catherine Wolf IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NY IBM Thomas J. Watson Research Center, 30 Saw Mill River Rd., Hawthorne, NYView Profile Authors Info & Claims CSCW '98: Proceedings of the 1998 ACM conference on Computer supported cooperative workNovember 1998 Pages 99–109https://doi.org/10.1145/289444.289484Published:01 November 1998 31citation820DownloadsMetricsTotal Citations31Total Downloads820Last 12 Months3Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

To model cache behavior in a multiprogramming environment, it is necessary to know the number of distinct lines referenced in an execution interval. The stack growth function (SGF) relates the mean number of references (or instructions) to the number of distinct lines referenced; it can be viewed as the inverse function of the mean working set size. A fast, one-pass algorithm to compute the SGF for a given referenced string is presented. SGFs measured for some 40 real programs show that a simple exponential model fits the SGFs reasonably well over a range of execution intervals. Parameters of the inverse exponential model are presented for several program mixes and cache line sizes of 32 and 64 bytes. Separate instruction and data SGFs also are examined, and execution interval distribution effects are considered.>

The "hit ratio" of a high-speed buffer (cache) depends on the "locality" of memory references. However, locality of reference is disturbed and the hit ratio decreases whenever a task switch occurs. This performance degradation can be minimized if "locality of task switching," the tendency for a small set of favored tasks to be frequently executed, exists and the cache is organized in such a way that it can hold blocks (lines) of multiple tasks. Locality of task switching and locality of memory references in individual tasks exhibit overall locality of memory references at a system level. This paper addresses the following questions. Does locality of task switching really exist? How can it be modeled? Task switching in IBM operating system/virtual storage with multiple virtual storage (OS/VS2 MVS) was measured using event traces for three different workloads to show that locality of task switching actually exists in MVS. Two different models of task switching are proposed. These models can be incorporated into cache multitasking models to predict more accurately the misses in real computer systems. A key parameter of these models is the task execution interval; measurements of execution intervals for the workloads used in the paper are presented and discussed.

Loops are directly defined on an instruction execution string rather than on a directed graph model of a program derived from the source code. A simple efficient algorithm to find the loops and subloops thus defined is presented.

Instruction mixes such as the Gibson mix have been used for a long time as workload models for CPU's. However, since an instruction mix does not indicate the order of instruction execution, it is not suitable for the performance evaluation of advanced computers which employ pipelining. Instruction sequences are proposed as a generalization of instruction mixes. The instruction representation does not consist of a fixed number of instructions, but rather it is defined to be a string of consecutively executed instructions terminated by the instructions which potentially interrupt sequential instruction streams. Statistics on sequences have been collected from traces of problem state programs. It was found that there are not many distinct sequences and that they are short. The average sequence length may be less than seven instructions. A small subset of the distinct sequences, say less than 20 percent, accounts for more than 95 percent of all executed instructions.