Search

Dynamic radio resource management (RRM) in wireless networks presents significant challenges, particularly in the context of Radio Access Network (RAN) slicing. This technology, crucial for catering to varying user requirements, often grapples with complex optimization scenarios. Existing Reinforcement Learning (RL) approaches, while achieving good performance in RAN slicing, typically rely on online algorithms or behavior cloning. These methods necessitate either continuous environmental interactions or access to high-quality datasets, hindering their practical deployment. Towards addressing these limitations, this paper introduces offline RL to solving the RAN slicing problem, marking a significant shift towards more feasible and adaptive RRM methods. We demonstrate how offline RL can effectively learn near-optimal policies from sub-optimal datasets, a notable advancement over existing practices. Our research highlights the inherent flexibility of offline RL, showcasing its ability to adjust policy criteria without the need for additional environmental interactions. Furthermore, we present empirical evidence of the efficacy of offline RL in adapting to various service-level requirements, illustrating its potential in diverse RAN slicing scenarios., Comment: 9 pages. 6 figures

Convolutional Sparse Coding (CSC) has been attracting more and more attention in recent years, for making full use of image global correlation to improve performance on various computer vision applications. However, very few studies focus on solving CSC based image Super-Resolution (SR) problem. As a consequence, there is no significant progress in this area over a period of time. In this paper, we exploit the natural connection between CSC and Convolutional Neural Networks (CNN) to address CSC based image SR. Specifically, Convolutional Iterative Soft Thresholding Algorithm (CISTA) is introduced to solve CSC problem and it can be implemented using CNN architectures. Then we develop a novel CSC based SR framework analogy to the traditional SC based SR methods. Two models inspired by this framework are proposed for pre-/post-upsampling SR, respectively. Compared with recent state-of-the-art SR methods, both of our proposed models show superior performance in terms of both quantitative and qualitative measurements., Comment: 10 pages, 12 figures, 3 tables

We propose a simple yet effective model for Single Image Super-Resolution (SISR), by combining the merits of Residual Learning and Convolutional Sparse Coding (RL-CSC). Our model is inspired by the Learned Iterative Shrinkage-Threshold Algorithm (LISTA). We extend LISTA to its convolutional version and build the main part of our model by strictly following the convolutional form, which improves the network's interpretability. Specifically, the convolutional sparse codings of input feature maps are learned in a recursive manner, and high-frequency information can be recovered from these CSCs. More importantly, residual learning is applied to alleviate the training difficulty when the network goes deeper. Extensive experiments on benchmark datasets demonstrate the effectiveness of our method. RL-CSC (30 layers) outperforms several recent state-of-the-arts, e.g., DRRN (52 layers) and MemNet (80 layers) in both accuracy and visual qualities. Codes and more results are available at https://github.com/axzml/RL-CSC., Comment: 10 pages, 8 figures, 4 tables

A critical challenge for wireless communications in the millimeter wave (mmWave) bands is blockage. MmWave signals suffer significant penetration losses from many common materials and objects, and small changes in the position of obstacles in the environment can cause large variations in the channel quality. This paper provides a measurement-based study of the effects of human blockage on an end-to-end application over a mmWave cellular link. A phased array system is used to measure the channel in multiple directions almost simultaneously in a realistic indoor scenario. The measurements are integrated into a detailed ns-3 simulation that models both the latest 3GPP New Radio beam search procedure as well as the internet protocol stack. The measurement-based simulation illustrates how recovery from blockage depends on the path diversity and beam search.

The vast available spectrum in the millimeter wave (mmWave) bands offers the possibility of multi-Gbps data rates for fifth generation (5G) cellular networks. However, mmWave capacity can be highly intermittent due to the vulnerability of mmWave signals to blockages and delays in directional searching. Such highly variable links present unique challenges for adaptive control mechanisms in transport layer protocols and end-to-end applications. This paper considers the fundamental question of whether TCP - the most widely used transport protocol - will work in mmWave cellular systems. The paper provides a comprehensive simulation study of TCP considering various factors such as the congestion control algorithm, including the recently proposed TCP BBR, edge vs. remote servers, handover and multi- connectivity, TCP packet size and 3GPP-stack parameters. We show that the performance of TCP on mmWave links is highly dependent on different combinations of these parameters, and identify the open challenges in this area., Comment: 7 pages, 4 figures, 2 tables. To be published in the IEEE Communication Magazine

TCP is the most widely used transport protocol in the internet. However, it offers suboptimal performance when operating over high bandwidth mmWave links. The main issues introduced by communications at such high frequencies are (i) the sensitivity to blockage and (ii) the high bandwidth fluctuations due to Line of Sight (LOS) to Non Line of Sight (NLOS) transitions and vice versa. In particular, TCP has an abstract view of the end-to-end connection, which does not properly capture the dynamics of the wireless mmWave link. The consequence is a suboptimal utilization of the available resources. In this paper we propose a TCP proxy architecture that improves the performance of TCP flows without any modification at the remote sender side. The proxy is installed in the Radio Access Network, and exploits information available at the gNB in order to maximize throughput and minimize latency., Comment: 7 pages, 6 figures, 2 tables, presented at the 2017 51st Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, 2017

Due to its potential for multi-gigabit and low latency wireless links, millimeter wave (mmWave) technology is expected to play a central role in 5th generation cellular systems. While there has been considerable progress in understanding the mmWave physical layer, innovations will be required at all layers of the protocol stack, in both the access and the core network. Discrete-event network simulation is essential for end-to-end, cross-layer research and development. This paper provides a tutorial on a recently developed full-stack mmWave module integrated into the widely used open-source ns--3 simulator. The module includes a number of detailed statistical channel models as well as the ability to incorporate real measurements or ray-tracing data. The Physical (PHY) and Medium Access Control (MAC) layers are modular and highly customizable, making it easy to integrate algorithms or compare Orthogonal Frequency Division Multiplexing (OFDM) numerologies, for example. The module is interfaced with the core network of the ns--3 Long Term Evolution (LTE) module for full-stack simulations of end-to-end connectivity, and advanced architectural features, such as dual-connectivity, are also available. To facilitate the understanding of the module, and verify its correct functioning, we provide several examples that show the performance of the custom mmWave stack as well as custom congestion control algorithms designed specifically for efficient utilization of the mmWave channel., Comment: 25 pages, 16 figures, submitted to IEEE Communications Surveys and Tutorials (revised Jan. 2018)

Communications at mmWave frequencies will be a key enabler of the next generation of cellular networks, due to the multi-Gbps rate that can be achieved. However, there are still several problems that must be solved before this technology can be widely adopted, primarily associated with the interplay between the variability of mmWave links and the complexity of mobile networks. An end-to-end network simulator represents a great tool to assess the performance of any proposed solution to meet the stringent 5G requirements. Given the criticality of channel propagation characteristics at higher frequencies, we present our implementation of the 3GPP channel model for the 6-100 GHz band for the ns-3 end-to-end 5G mmWave module, and detail its associated MIMO beamforming architecture.

Due to massive available spectrum in the millimeter wave (mmWave) bands, cellular systems in these frequencies may provides orders of magnitude greater capacity than networks in conventional lower frequency bands. However, due to high susceptibility to blocking, mmWave links can be extremely intermittent in quality. This combination of high peak throughputs and intermittency can cause significant challenges in end-to-end transport-layer mechanisms such as TCP. This paper studies the particularly challenging problem of bufferbloat. Specifically, with current buffering and congestion control mechanisms, high throughput-high variable links can lead to excessive buffers incurring long latency. In this paper, we capture the performance trends obtained while adopting two potential solutions that have been proposed in the literature: Active queue management (AQM) and dynamic receive window. We show that, over mmWave links, AQM mitigates the latency but cannot deliver high throughput. The main reason relies on the fact that the current congestion control was not designed to cope with high data rates with sudden change. Conversely, the dynamic receive window approach is more responsive and therefore supports higher channel utilization while mitigating the delay, thus representing a viable solution.

The millimeter wave (mmWave) bands are likely to play a significant role in next generation cellular systems due to the possibility of very high throughput thanks to the availability of massive bandwidth and high-dimensional antennas. Especially in Non-Line-of-Sight conditions, significant variations in the received RF power can occur as a result of the scattering from nearby building and terrain surfaces. Scattering objects come and go as the user moves through the local environment. At the higher end of the mmWave band, rough surface scatter generates cluster-based small-scale fading, where signal levels can vary by more than 20 dB over just a few wavelengths. This high level of channel variability may present significant challenges for congestion control. Using our recently developed end-to-end mmWave ns3-based framework, this paper presents the first performance evaluation of TCP congestion control in next-generation mmWave networks. Importantly, the framework can incorporate detailed models of the mmWave channel, beam- forming and tracking algorithms, and builds on statistical channel models derived from real measurements in New York City, as well as detailed ray traces.

The growing demand for ubiquitous mobile data services along with the scarcity of spectrum in the sub-6 GHz bands has given rise to the recent interest in developing wireless systems that can exploit the large amount of spectrum available in the millimeter wave (mmWave) frequency range. Due to its potential for multi-gigabit and ultra-low latency links, mmWave technology is expected to play a central role in 5th Generation (5G) cellular networks. Overcoming the poor radio propagation and sensitivity to blockages at higher frequencies presents major challenges, which is why much of the current research is focused at the physical layer. However, innovations will be required at all layers of the protocol stack to effectively utilize the large air link capacity and provide the end-to-end performance required by future networks. Discrete-event network simulation will be an invaluable tool for researchers to evaluate novel 5G protocols and systems from an end-to-end perspective. In this work, we present the first-of-its-kind, open-source framework for modeling mmWave cellular networks in the ns-3 simulator. Channel models are provided along with a configurable physical and MAC-layer implementation, which can be interfaced with the higher-layer protocols and core network model from the ns-3 LTE module to simulate end-to-end connectivity. The framework is demonstrated through several example simulations showing the performance of our custom mmWave stack.

The IMT 2020 requirements of 20 Gbps peak data rate and 1 millisecond latency present significant engineering challenges for the design of 5G cellular systems. Use of the millimeter wave (mmWave) bands above 10 GHz --- where vast quantities of spectrum are available --- is a promising 5G candidate that may be able to rise to the occasion. However, while the mmWave bands can support massive peak data rates, delivering these data rates on end-to-end service while maintaining reliability and ultra-low latency performance will require rethinking all layers of the protocol stack. This papers surveys some of the challenges and possible solutions for delivering end-to-end, reliable, ultra-low latency services in mmWave cellular systems in terms of the Medium Access Control (MAC) layer, congestion control and core network architecture.

The millimeter-wave (mmWave) frequencies have attracted considerable attention for fifth generation (5G) cellular communication as they offer orders of magnitude greater bandwidth than current cellular systems. However, the medium access control (MAC) layer may need to be significantly redesigned to support the highly directional transmissions, ultra-low latencies and high peak rates expected in mmWave communication. To address these challenges, we present a novel mmWave MAC layer frame structure with a number of enhancements including flexible, highly granular transmission times, dynamic control signal locations, extended messaging and ability to efficiently multiplex directional control signals. Analytic formulae are derived for the utilization and control overhead as a function of control periodicity, number of users, traffic statistics, signal-to-noise ratio and antenna gains. Importantly, the analysis can incorporate various front-end MIMO capability assumptions -- a critical feature of mmWave. Under realistic system and traffic assumptions, the analysis reveals that the proposed flexible frame structure design offers significant benefits over designs with fixed frame structures similar to current 4G long-term evolution (LTE). It is also shown that fully digital beamforming architectures offer significantly lower overhead compared to analog and hybrid beamforming under equivalent power budgets., Comment: Submitted to IEEE Transactions for Wireless Communications

The increasing demand of data, along with the spectrum scarcity, are motivating a urgent shift towards exploiting new bands. This is the main reason behind identifying mmWaves as the key disruptive enabling technology for 5G cellular networks. Indeed, utilizing new bands means facing new challenges; in this context, they are mainly related to the radio propagation, which is shorter in range and more sensitive to obstacles. The resulting key aspects that need to be taken into account when designing mmWave cellular systems are directionality and link intermittency. The lack of network level results motivated this work, which aims at providing the first of a kind open source mmWave framework, based on the network simulator ns-3. The main focus of this work is the modeling of customizable channel, physical (PHY) and medium access control (MAC) layers for mmWave systems. The overall design and architecture of the model are discussed in details. Finally, the validity of our proposed framework is corroborated through the simulation of a simple scenario.