Kaikki aineistot
Lisää
This thesis evaluates Automatic Test Framework’s behavior when scheduling test case to run on real hardware sets. The company has a limited amount of test hardware sets for developers to manually choose to perform different test cases. Each test case requires a different hardware setup and a different execution time. Ideally, a test should be performed on a minimal required hardware set that satisfies the test requirements. As a result, it complicates finding the most optimal way of scheduling the test execution manually. Furthermore, test hardware is expensive, and the company wants to invest more on software solution to utilize test hardware to serve testing purpose instead of buying more hardware. Hence, a solution, which runs on a web application, is designed to automatically take all the mentioned information into consideration and provide the most optimal schedule for the test execution planning. The result of this work helps to reduce manual work and time spent on the testing phase.
The ever-increasing pace of neural network (NN) based solutions for computer vision tasks is making them one of the main consumers of digital images nowadays. This raises the question of whether the traditional human-oriented image codecs, or the adapted version of these codecs for the machine-targeted use cases are efficient enough for the massive amount of image data generated every day for both humans and machines. This thesis explores the abilities of the image codecs that are designed specifically only for machine-consumption. To the best of the student’s knowledge, this is the first end-to-end learned machine-oriented image codec proposal. It presents an end-to-end framework for designing NN-based image codecs for machines, as well as a set of training strategies that address the delicate problem of balancing competing losses in multi-task training, namely image distortion loss, rate loss, and computer vision task losses. The experimental results show the superior coding efficiency of the proposed codecs in comparison with the current state-of-the-art standard VVC/H.266 on object detection and instance segmentation, achieving -37.87% and -32.90% BD-rate gain, respectively while being extremely fast thanks to its compact size. These results also serve as a proof-of-concept for a new approach to Image coding for machines.
The topic of this report is “The usage of React Native in managing and operating the system in Logistic Industry”. This thesis is going to introduce the fundamentals of Mobile Development and Logistic Industry. Furthermore, the problems and challenges in this business will be presented along with the solution when using React Native. With the new features and the capability of integration of new technologies, React Native can be used effectively in improving warehouse management systems performance. Inheriting from Javascript, one of the most well-known programming languages, React Native was developed by Facebook company for implementing cross-platform application in both Android and iOS devices. React Native has a similar concept with React, the web development tool; but instead of using web components for developing the user interfaces, React Native uses the built-in native components for the mobile platforms. After being released in 2015, both React and React Native have gained popularity among software and mobile community, and they are being used in a wide range of areas of daily life, especially in Logistics and Warehouse Industry. A mobile application is going to be developed and launched using React Native for idea.invest, a Logistics and Software Solutions company in Hamburg, Germany.
The State Audit of Vietnam (abbreviated as SAV) was established in the country's renewal process and the process of reforming the national administration. Restrictions on state budget estimates impose a duty on the SAV to participate deeply and broadly in the preparation and appraisal of state budget. As stipulated in Law on State Audit, the SAV has enough legal basis to conduct the audit of state budget estimates. In fact, the SAV has made many contributions to the management and administration of state budget of levels of government. It also helps the budgetary supervision of NA and the People's Councils at all levels, including the provision of information and documents to NA on annual estimates of state budget. However, the quality of auditing was inadequate, not meeting requirements because this is a new task that has not preceded before. The SAV lacked both theory and practical experiences to audit state budget estimates. Therefore, I chose the topic "Evaluation of organizing state budget estimates audit of Vietnam". The objective of this study is to evaluate what is required of the management and operation of the State audit when auditing state budget estimates.
The combination of business with information technology has existed for a long time. Nowadays, we hear the word "digitalization" in every business because digitizing data helps to improve business efficiency, reduce costs, and improve employee productivity in the business. This has long been no longer a controversial issue but has been proven by countless businesses. In the past, to new companies, building platforms, good tools, and infrastructure to enable digitalization of data was not an easy task due to human resource and costs constraints. However, with developing of current technology, this has been simplified, and every business can do it easily. Therefore, in the thesis, we would like to introduce benefits of technologies in business and how technologies bring the result on business. The combination will bring convenient to customers, help company are easier in managing and keeping loyalty customer. We also introduce cutting-edge, typical, and easy-to-use technologies when building web-based software, as well as phone applications, for digitizing data in businesses.
Abstract We proposed several energy-efficient resource allocation algorithms for the downlink of an orthogonal frequency-division-multiple-access (OFDMA) based femtocell heterogeneous networks (HetNets). Heterogeneous QoS and fairness in rate are investigated in the proposed resource allocation problem. A dense deployment of femtocells in the coverage area of a central macrocell is considered and energy usage of both femtocell and macrocell users are optimized simultaneously. We aim to maximize the weighted sum of the individual energy efficiencies (WSEEMax) and the network energy efficiency (NEEMax) while satisfying the following: (1) minimum throughput for delay-sensitive (DS) users, (2) fairness constraint for delay-tolerant (DT) users, (3) required constraints of OFDMA systems. The problem is formulated in three different forms: mixed 0—1 integer programming formulation, time-sharing formulation and sparsity-inducing formulation. The proposed resource block (RB) and power optimization problems are combinatorial and highly non-convex due to the fractional form of the objective function, the integer constraint of OFDMA RBs and non-affine fairness. We adopt the successive convex approximation (SCA) approach and transform the problems into a sequence of convex subproblems. With the proposed algorithms, we show that the overall joint RB and power allocation schemes converge to suboptimal solutions. Numerical examples confirm the merits of the proposed algorithms.
Over recent years, deep learning-based computer vision systems have been applied to images at an ever-increasing pace, oftentimes representing the only type of consumption for those images. Given the dramatic explosion in the number of images generated per day, a question arises: how much better would an image codec targeting machine-consumption perform against state-of-the-art codecs targeting human-consumption? In this paper, we propose an image codec for machines which is neural network (NN) based and end-to-end learned. In particular, we propose a set of training strategies that address the delicate problem of balancing competing loss functions, such as computer vision task losses, image distortion losses, and rate loss. Our experimental results show that our NN-based codec outperforms the state-of-the-art Versatile Video Coding (VVC) standard on the object detection and instance segmentation tasks, achieving -37.87% and -32.90% of BD-rate gain, respectively, while being fast thanks to its compact size. To the best of our knowledge, this is the first end-to-end learned machine-targeted image codec.
Abstract We consider a multigroup multicast cell-free multiple-input multiple-output (MIMO) downlink system with short-term power constraints. In particular, the normalized conjugate beamforming scheme is adopted at each access point (AP) to keep the downlink power strictly under the power budget regardless of small scale fading. In the considered scenario, APs multicast signals to multiple groups of users whereby users in the same group receive the same message. Under this setup, we are interested in maximizing the minimum achievable rate of all groups, commonly known as the max-min fairness problem, which has not been studied before in this context. To solve the considered problem, we first present a bisection method which in fact has been widely used in previous studies for cell-free massive MIMO, and then propose an accelerated projected gradient (APG) method. We show that the proposed APG method outperforms the bisection method requiring lesser run time while still achieving the same objective value. Moreover, the considered power control scheme provides significantly improved performance and more fairness among the users compared to the equal power allocation scheme.
Abstract We study the problem of energy efficiency maximization (EEmax) with joint beamforming and subarray selection, by taking into account the non-linear power amplifier (PA) efficiency in a multi-user multiple-input single-output system. The subarray selection problem is formulated using the concept of perspective formulation with additional penalty term in the objective function. To tackle the resulting challenging mixed-Boolean non-convex optimization problem, we rely on continuous relaxation and successive convex approximation framework where a convex problem is solved in each iteration. Numerical results demonstrate the achieved energy efficiency gains of the subarray selection and show that non-linear PA efficiency has a significant impact on the optimization. We also observe that on contrast to using linear PA efficiency model, the non-linear PA efficiency model yields the fact that it is better to stay silent rather than transmit with very low transmit power.
Abstract We investigate the coordinated multi-point noncoherent joint transmission (JT) in dense small cell networks. The goal is to design beamforming vectors for macro cell and small cell base stations (BSs) such that the weighted sum rate of the system is maximized, subject to a total transmit power at individual BSs. The optimization problem is inherently nonconvex and intractable, making it difficult to explore the full potential performance of the scheme. To this end, we first propose an algorithm to find a globally optimal solution based on the generic monotonic branch reduce and bound optimization framework. Then, for a more computationally efficient method, we adopt the inner approximation (InAp) technique to efficiently derive a locally optimal solution, which is numerically shown to achieve near-optimal performance. In addition, for decentralized networks such as those comprising of multi-access edge computing servers, we develop an algorithm based on the alternating direction method of multipliers, which distributively implements the InAp-based solution. Our main conclusion is that the noncoherent JT is a promising transmission scheme for dense small cell networks, since it can exploit the densitification gain, outperforms the coordinated beamforming, and is amenable to distributed implementation.
Abstract We consider downlink transmission whereby a multiantenna base station simultaneously transmits data to multiple single-antenna users. We focus on slow flat fading channel where the channel state information is imperfect, the channel estimation error is unbounded and its statistics are known. The aim is to design beamforming vectors such that the sum rate is maximized under the constraints on probability of successful transmission for each user and maximum transmit power. The optimization problem is intractable due to the chance constraints. To this end, we propose an efficient solution drawn upon stochastic optimization. In particular, we first use the step function and its smooth approximation to get an approximate nonconvex stochastic program of the considered problem. We then develop an iterative procedure to solve the stochastic program based on the stochastic successive convex approximation framework. The numerical results show that the proposed solution can achieve remarkable sum rate gains compared to the conventional one.
Abstract Intelligent reflecting surfaces (IRSs) have shown huge advantages in many potential use cases and thus have been considered a promising candidate for next-generation wireless systems. In this paper, we consider an IRS-assisted multigroup multicast (IRS-MGMC) system in a multiple-input single-output (MISO) scenario, for which the related existing literature is rather limited. In particular, we aim to jointly design the transmit beamformers and IRS phase shifts to maximize the sum rate of the system under consideration. In order to obtain a numerically efficient solution to the formulated non-convex optimization problem, we propose an alternating projected gradient (APG) method where each iteration admits a closed-form and is shown to be superior to a known solution that is derived from the majorization-minimization (MM) method in terms of both achievable sum rate and required complexity, i.e., run time. In particular, we show that the complexity of the proposed APG method grows linearly with the number of IRS tiles, while that of the known solution in comparison grows with the third power of the number of IRS tiles. The numerical results reported in this paper extend our understanding on the achievable rates of large-scale IRS-assisted multigroup multicast systems.
Abstract We study the problem of designing transmit beamformer for maximizing the weight sum rate of a multi-user (MU)-multiple-input single-output (MISO) interference broadcast channel (IBC) with individual quality-of-service (QoS) constraints. The considered problem is known to be nonconvex and NP-hard indeed, and most of existing high-performance solutions are based on the centralized method. In this paper, we propose a distributed approach for the weighted sum rate maximization (WSRM) problem with QoS constraints by combining successive convex approximation (SCA) framework and the alternating directions method of multipliers (ADMM) technique. More specifically, the proposed algorithm extends a current centralized solution, where the SCA is used to arrive at an approximate convex problem at each step of the iterative procedure. The idea is that we apply the ADMM technique to solve the convex problem of the SCA based subproblem in a distributed manner. We also discuss some heuristic ways to accelerate the convergence rate of the proposed algorithm. Numerical simulations are provided to compare different models for both centralized and distributed algorithms.
Abstract This paper proposes energy-efficient coordinated beamforming strategies for multicell multiuser multiple-input single-output system. We consider a practical power consumption model, where part of the consumed power depends on the base station or user specific data rates due to coding, decoding, and backhaul. This is different from the existing approaches where the base station power consumption has been assumed to be a convex or linear function of the transmit powers. Two optimization criteria are considered, namely network energy efficiency maximization and weighted sum energy efficiency maximization. We develop successive convex approximation-based algorithms to tackle these difficult nonconvex problems. We further propose decentralized implementations for the considered problems, in which base stations perform parallel and distributed computation based on local channel state information and limited backhaul information exchange. The decentralized approaches admit closed-form solutions and can be implemented without invoking a generic external convex solver. We also show an example of the pilot contamination effect on the energy efficiency using a heuristic pilot allocation strategy. The numerical results are provided to demonstrate that the rate dependent power consumption has a large impact on the system energy efficiency, and, thus, has to be taken into account when devising energy-efficient transmission strategies. The significant gains of the proposed algorithms over the conventional low-complexity beamforming algorithms are also illustrated.
Abstract We study the problem of designing multicast precoders for multiple groups with the objective of minimizing total transmit power under certain guaranteed quality-of-service (QoS) requirements. To avail both spatial and frequency diversity, we consider a multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) system. The problem of interest is in fact a nonconvex quadratically constrained quadratic program (QCQP) for which the prevailing semidefinite relaxation (SDR) technique is inefficient for at least two reasons. At first, the relaxed problem cannot be equivalently reformulated as a semidefinite programming (SDP). Secondly, even if the relaxed problem is solved, the so-called randomization procedure should be used to generate a high quality feasible solution to the original QCQP. However, such a randomization procedure is difficult in the considered system model. To overcome these shortcomings, we adopt successive convex approximation (SCA) framework in this paper to find beamformers directly. The proposed method not only avoids the randomization procedure mentioned above but also requires lower computational complexity compared to the SDR approach. Numerical experiments are carried out to demonstrate the effectiveness of the proposed algorithm.
Abstract The concept of the wireless dynamic network architecture (DNA) stands for a system design, which allows terminals to convert into temporary access points (APs) when necessary. In this paper, we propose a framework to solve the problem of load balancing in DNA. Particularly, the user association in DNAs is optimized to minimize the number of active APs and the network cost in terms of tradeoff between power and load, while ensuring users’ quality of service (QoS). In general, such a problem is a non-convex mixed integer nonlinear program in the sense that its continuous relaxation is a non-convex problem. To solve this optimization, we use the standard continuous relaxation method and approximate the relaxed problem by a series of second order cone programs with the aid of successive convex approximation (SCA) framework. Numerical results show that the proposed algorithm converges within a few iterations and jointly minimizes the network cost and the number of APs in the network.
Abstract We develop a novel technique to accelerate minorization-maximization (MM) procedure for the non-orthogonal multiple access (NOMA) weighted sum rate maximization problem. Specifically, we exploit the Lipschitz continuity of the gradient of the objective function to adaptively update the MM algorithm. With fewer additional analysis variables and low complexity second-order cone program (SOCP) to solve in each iteration of the MM algorithm, the proposed approach converges quickly at a small computational cost. By numerical simulation results, our algorithm is shown to greatly outperform known solutions in terms of achieved sum rates and computational complexity.
Abstract To mitigate the computational power gap between the network core and edges, mobile edge computing (MEC) is poised to play a fundamental role in future generations of wireless networks. In this correspondence, we consider a non-orthogonal multiple access (NOMA) transmission model to maximize the worst task to be offloaded among all users to the network edge server. A provably convergent and efficient algorithm is developed to solve the considered non-convex optimization problem for maximizing the minimum number of offloaded bits in a multi-user NOMA-MEC system. Compared to the approach of optimized orthogonal multiple access (OMA), for given MEC delay, power and energy limits, the NOMA-based system considerably outperforms its OMA-based counterpart in MEC settings. Numerical results demonstrate that the proposed algorithm for NOMA-based MEC is particularly useful for delay sensitive applications.
Abstract This paper studies energy-efficient joint transmit and receive beamforming in multi-cell multi-user multiple-input multiple-output systems. We consider conventional network energy efficiency metric where the users can receive unicasting streams in addition to the group-specific common multicasting streams which have certain rate constraints. The goal is to use the transmission resources more efficiently to improve the energy efficiency, when the users are equipped with multiple antennas. Numerical results show the achieved energy efficiency gains by using the additional degrees of freedom of the multicasting transmission to private message unicasting.
Abstract This paper studies energy-efficient joint coordinated beamforming and antenna selection in multi-cell multi-user multigroup multicast multiple-input single-output systems. We focus on interference-limited scenarios, e.g., when the number of radio frequency (RF) chains is of the same order as the number of multicasting groups. To tackle the interference, we exploit rate-splitting to divide the group messages into common and group-specific sub-messages. We propose a per-cell rate-splitting approach, where the common message is locally designed to be decoded by the in-cell users, while treated as noise by the out-cell users. We consider the case where the number of RF chains is smaller than that of antennas, and consider a switching architecture, that is, the antenna selection is employed to choose the best antennas for transmission. Numerical results illustrate the potential of the proposed approach to significantly improve the energy efficiency in the interference-limited regime.
Abstract This paper aims at guaranteeing the achievable energy efficiency (EE) fairness in a multicell multiuser multiple-input single-output downlink system. The design objective is to maximize the minimum EE among all base stations (BSs) subject to per-BS power constraints. This results in a max-min fractional program and as such is difficult to solve in general. Our goal is to develop decentralized algorithms for the max-min EE problem based on combining the successive convex approximation (SCA) framework and the alternating direction method of multipliers (ADMMs). Specifically, leveraging the SCA principle, we iteratively approximate the nonconvex design problem by a sequence of convex programs for which two decentralized algorithms are then proposed. In the first approach, the convex program obtained at each step of the SCA procedure is solved optimally by allowing the BSs to exchange the required information until the ADMM converges. The convergence of the first method is analytically guaranteed but the amount of backhaul signaling can be noticeable in some realistic settings. To reduce the backhaul overhead, the second method performs an abstract version of the ADMM where only one variables update is carried out. Numerical results are provided to demonstrate the effectiveness of the two proposed decentralized algorithms.
Abstract This paper considers a downlink transmission of cloud radio access network (C-RAN) in which precoded baseband signals at a common baseband unit are compressed before being forwarded to radio units (RUs) through limited fronthaul capacity links. We investigate the joint design of precoding, multivariate compression and RU-user selection which maximizes the energy efficiency of downlink C-RAN networks. The considered problem is inherently a rank-constrained mixed Boolean nonconvex program for which a globally optimal solution is difficult and computationally expensive to find. In order to derive practically appealing solutions, we invoke some useful relaxation and transformation techniques to arrive at a more tractable (but still nonconvex) continuous program. To solve the relaxation problem, we propose an iterative procedure based on DC algorithms which is provably convergent. Numerical results demonstrate the superior of the proposed solution in terms of achievable energy efficiency compared to existing schemes.
Abstract We consider a multi-pair amplify-and-forward relay network where the energy-constrained relays adopting the time-switching protocol harvest energy from the radio-frequency signals transmitted by the users for assisting user data transmission. Both one-way and two-way relaying techniques are investigated. Aiming at energy efficiency (EE) fairness among the user pairs, we construct an energy consumption model incorporating rate-dependent signal processing power, the dependence on output power level of power amplifiers’ efficiency, and nonlinear energy harvesting (EH) circuits. Then, we formulate the max-min EE fairness problems in which the data rates, users’ transmit power, relays’ processing coefficient, and EH time are jointly optimized under the constraints on the quality of service and users’ maximum transmit power. To achieve efficient suboptimal solutions to these nonconvex problems, we devise monotonic descent algorithms based on the inner approximation (IA) framework, which solve a second-order-cone program in each iteration. To further simplify the designs, we propose an approach combining IA and zero-forcing beamforming, which eliminates inter-pair interference and reduces the numbers of variables and required iterations. Finally, extensive numerical results are presented to validate the proposed approaches. More specifically, the results demonstrate that ignoring the realistic aspects of power consumption might degrade the performance remarkably, and jointly designing parameters involved could significantly enhance the EE.
Abstract We study downlink of multiantenna cloud radio access networks with finite-capacity fronthaul links. The aim is to propose joint designs of beamforming and remote radio head (RRH)-user association, subject to constraints on users’ quality-of-service, limited capacity of fronthaul links and transmit power, to maximize the system energy efficiency. To cope with the limited-capacity fronthaul we consider the problem of RRH-user association to select a subset of users that can be served by each RRH. Moreover, different to the conventional power consumption models, we take into account the dependence of the baseband signal processing power on the data rate, as well as the dynamics of the efficiency of power amplifiers. The considered problem leads to a mixed binary integer program which is difficult to solve. Our first contribution is to derive a globally optimal solution for the considered problem by customizing a discrete branch-reduce-and-bound approach. Since the global optimization method requires a high computational effort, we further propose two suboptimal solutions able to achieve the near optimal performance but with much reduced complexity. To this end, we transform the design problem into continuous (but inherently nonconvex) programs by two approaches: penalty and l0 -approximation methods. These resulting continuous nonconvex problems are then solved by the successive convex approximation framework. Numerical results are provided to evaluate the effectiveness of the proposed approaches.
Abstract We consider a secure transmission including a transmitter, a receiver and an eavesdropper, each being equipped with multiple antennas. The aim is to develop a low-complexity and scalable method to find a globally optimal solution to the problem of secrecy rate maximization under a total power constraint at the transmitter. In principle, the original formulation of the problem is nonconvex. However, it can be equivalently translated into finding a saddle point of a minimax convex-concave program. An existing approach finds the saddle point using the Newton method, whose computational cost increases quickly with the number of transmit antennas, making it unsuitable for large scale antenna systems. To this end, we propose an iterative algorithm based on alternating optimization, which is guaranteed to converge to a saddle point, and thus achieves a globally optimal solution to the considered problem. In particular, each subproblem of the proposed iterative method admits a closed-form solution. We analytically show that the iteration cost of our proposed method is much cheaper than that of the known solution. As a result, numerical results demonstrate that the proposed method remarkably outperforms the existing one in terms of the overall run time.
Abstract This paper considers a multicell downlink channel in which multiple base stations (BSs) cooperatively serve users by jointly precoding shared data transported from a central processor over limited-capacity backhaul links. We jointly design the beamformers and BS-user link selection so as to maximize the sum rate subject to user-specific signal-to-interference-noise (SINR) requirements, per-BS backhaul capacity and per-BS power constraints. As existing solutions for the considered problem are suboptimal and their optimality remains unknown due to the lack of globally optimal solutions, we characterized this gap by proposing an globally optimal algorithm for the problem of interest. Specifically, the proposed method is customized from a generic framework of a branch and bound algorithm applied to discrete monotonic optimization. We show that the proposed algorithm converges after a finite number of iterations, and can serve as a benchmark for existing suboptimal solutions and those that will be developed for similar contexts in the future. In this regard, we numerically compare the proposed optimal solution to a current state-of-the-art, which show that this suboptimal method only attains 70% to 90% of the optimal performance.
Abstract We consider downlink transmission of a fronthaul-constrained cloud radio access network. Our aim is to maximize the system sum data rate via jointly designing beamforming and user association. The problem is basically a mixed integer non-convex programs for which a global solution requires a prohibitively high computational effort. The focus is thus on efficient solutions capable of achieving the near optimal performance with low complexity. To this end, we transform the design problem into continuous programs by two approaches: penalty and sparse approximation methods. The resulting continuous nonconvex problems are then solved by the successive convex approximation framework. Numerical results indicate that the proposed methods are near-optimal, and outperform existing suboptimal methods in terms of achieved performances and computational complexity.
Abstract We consider uplink multiuser wireless communications systems, where the base station (BS) receiver is equipped with a large-scale antenna array and resolution adaptive analog-to-digital converters (ADCs). The aim is to maximize the energy efficiency (EE) at the BS subject to constraints on the users’ quality-of-service. The approach is to jointly optimize both the number of quantization bits at the ADCs and the on/off modes of the radio frequency (RF) processing chains. The considered problem is a discrete nonlinear program, the optimal solution of which is difficult to find. We develop an efficient algorithm based on the discrete branch-reduce-and-bound (DBRnB) framework. It finds the globally optimal solutions to the problem. In particular, we make some modifications, which significantly improve the convergence performance. The numerical results demonstrate that optimizing jointly the number of quantization bits and on/off mode can achieve remarkable EE gains compared to only optimizing the number of quantization bits.
Abstract Cloud radio access network (C-RAN) is an evolutionary radio network architecture in which a cloud-computing-based baseband (BB) signal-processing unit is shared among distributed low-cost wireless access points. This architecture offers a number of significant improvements over the traditional RANs, including better network scalability, spectral, and energy efficiency. As such C -RAN has been identified as one of the enabling technologies for the next-generation mobile networks. This chapter focuses on examining the energy-efficient transmission strategies of the C-RAN for cellular systems. In particular, we present optimization algorithms for the problem of transmit beamforming designs maximizing the network energy efficiency. In general, the energy efficiency maximization in C-RANs inherits the difficulty of resource allocation optimizations in interference-limited networks, i.e., it is an intractable non convex optimization problem. We first introduce a globally optimal method based on monotonic optimization (MO) to illustrate the optimal energy efficiency performance of the considered system. While the global optimization method requires extremely high computational effort and, thus, is not suitable for practical implementation, efficient optimization techniques achieving near -optimal performance are desirable in practice. To fulfill this gap, we present three low -complexity approaches based on the state-of-the-art local optimization framework, namely, successive convex approximation (SCA).
Abstract This paper studies the fairness of achievable energy efficiency (EE) in a multicell multiuser multiple-input single-output downlink. The objective is to maximize the minimum EE among all base stations (BSs) subject to per-BS power constraints. The resulting optimization problem is a max-min fractional program, and, thus, difficult to solve in general. Our goal is to develop a decentralized algorithm for the max-min EE problem which solves the problem locally. The idea behind the proposed method is to combine the framework of successive convex approximation (SCA) and alternative direction method of multipliers (ADMM). We transform the convex program obtained at each step of the SCA procedure into a form that lends itself to the ADMM. The resulting formulation is solved optimally by allowing the BSs to exchange the required information until the ADMM converges. In addition to further reduce the backhaul overhead, the proposed algorithm is modified to enhance the convergence speed. Numerical results are provided to demonstrate the effectiveness of the proposed algorithms.
Abstract A joint beamforming and remote radio head (RRH)-user association design for downlink of cloud radio access networks (CRANs) is considered. The aim is to maximize the system energy efficiency subject to constraints on users’ quality-of-service, capacity of front haul links and transmit power. Different to the conventional power consumption models, we embrace the dependence of baseband signal processing power on the data rate, and the dynamics of the power amplifiers’ efficiency. The considered problem is a mixed Boolean nonconvex program whose optimal solution is difficult to find. As our main contribution, we provide a discrete branch-reduce-and-bound (DBRnB) approach to solve the problem globally. We also make some modifications to the standard DBRnB procedure. Those remarkably improve the convergence performance. Numerical results are provided to confirm the validity of the proposed method.
Image coding for machines (ICM) aims at reducing the bitrate required to represent an image while minimizing the drop in machine vision analysis accuracy. In many use cases, such as surveillance, it is also important that the visual quality is not drastically deteriorated by the compression process. Recent works on using neural network (NN) based ICM codecs have shown significant coding gains against traditional methods; however, the decompressed images, especially at low bitrates, often contain checkerboard artifacts. We propose an effective decoder finetuning scheme based on adversarial training to significantly enhance the visual quality of ICM codecs, while preserving the machine analysis accuracy, without adding extra bitcost or parameters at the inference phase. The results show complete removal of the checkerboard artifacts at the negligible cost of −1.6% relative change in task performance score. In the cases where some amount of artifacts is tolerable, such as when machine consumption is the primary target, this technique can enhance both pixel-fidelity and feature-fidelity scores without losing task performance.
Image coding for machines (ICM) aims at reducing the bitrate required to represent an image while minimizing the drop in machine vision analysis accuracy. In many use cases, such as surveillance, it is also important that the visual quality is not drastically deteriorated by the compression process. Recent works on using neural network (NN) based ICM codecs have shown significant coding gains against traditional methods; however, the decompressed images, especially at low bitrates, often contain checkerboard artifacts. We propose an effective decoder finetuning scheme based on adversarial training to significantly enhance the visual quality of ICM codecs, while preserving the machine analysis accuracy, without adding extra bitcost or parameters at the inference phase. The results show complete removal of the checkerboard artifacts at the negligible cost of −1.6% relative change in task performance score. In the cases where some amount of artifacts is tolerable, such as when machine consumption is the primary target, this technique can enhance both pixel-fidelity and feature-fidelity scores without losing task performance.
Abstract Low latency communications is one of the key design targets in future wireless networks. We propose a queue aware algorithm to optimize resources guaranteeing low latency in multiple-input single-output (MISO) networks. Proposed system model is based on dynamic network architecture (DNA), where some terminals can be configured as temporary access points (APs) on demand when connected to the Internet. Therein, we jointly optimize the user-AP association and queue weighted sum rate of the network, subject to limitations of total transmit power of the APs and minimum delay requirements of the users. The user-AP association is viewed as finding a sparsity constrained solution to the problem of minimizing l q -norm of the difference between queue and service rate of users. Finally, the efficacy of the proposed algorithm in terms of network latency and its fast convergence are demonstrated using numerical experiments. Simulation results show that the proposed algorithm yields up to two-fold latency reductions compared to the state-of-the-art techniques.
Abstract This paper studies the energy efficiency and sum rate tradeoff for coordinated beamforming in multicell multiuser multigroup multicast multiple-input single-output systems. We first consider a conventional network energy efficiency maximization (EEmax) problem by jointly optimizing the transmit beamformers and antennas selected to be used in transmission. We also account for per-antenna maximum power constraints to avoid nonlinear distortion in power amplifiers and user-specific minimum rate constraints to guarantee certain service levels and fairness. To be energy efficient, transmit antenna selection is employed. It eventually leads to a mixed-Boolean fractional program. We then propose two different approaches to solve this difficult problem. The first solution is based on a novel modeling technique that produces a tight continuous relaxation. The second approach is based on sparsity-inducing method, which does not require the introduction of any Boolean variable. We also investigate the tradeoff between the energy efficiency and sum rate by proposing two different formulations. In the first formulation, we propose a new metric, that is, the ratio of the sum rate and the so-called weighted power. Specifically, this metric reduces to EEmax when the weight is 1, and to sum rate maximization when the weight is 0. In the other method, we treat the tradeoff problem as a multiobjective optimization for which a scalarization approach is adopted. Numerical results illustrate significant achievable energy efficiency gains over the method where the antenna selection is not employed. The effect of antenna selection on the energy efficiency and sum rate tradeoff is also demonstrated.
Machine-To-Machine (M2M) communication applications and use cases, such as object detection and instance segmentation, are becoming mainstream nowadays. As a consequence, majority of multimedia content is likely to be consumed by machines in the coming years. This opens up new challenges on efficient compression of this type of data. Two main directions are being explored in the literature, one being based on existing traditional codecs, such as the Versatile Video Coding (VVC) standard, that are optimized for human-Targeted use cases, and another based on end-To-end trained neural networks. However, traditional codecs have significant benefits in terms of interoperability, real-Time decoding, and availability of hardware implementations over end-To-end learned codecs. Therefore, in this paper, we propose learned post-processing filters that are targeted for enhancing the performance of machine vision tasks for images reconstructed by the VVC codec. The proposed enhancement filters provide significant improvements on the target tasks compared to VVC coded images. The conducted experiments show that the proposed post-processing filters provide about 45% and 49% Bjontegaard Delta Rate gains over VVC in instance segmentation and object detection tasks, respectively.
Wireless power transfer presents a robust method to provide a safe and convenient solution for powering biomedical implants. The implanted device is necessarily minimized conducive to the transfer efficiency degradation. This paper presents a wireless power transfer for retinal prosthesis using an artificial intelligent algorithm to optimize the power transfer efficiency with the minimizing implanted device dimension. A novel transmitter is designed based on the conformal strongly coupled magnetic resonator, which is proposed to attach to the glasses lens. An implanted receiver is optimized to obtain the highest efficiency with the minimal dimension of 9.7 x 9.7 x 1.9 mm3. The WPT system achieves approximately 17.5 % at the operating ISM frequency band of 433 MHz and a distance between transmitter and receiver of 20 mm.
Abstract This paper focuses on energy-efficient coordinated multi-point (CoMP) downlink in multi-antenna multi-cell wireless communications systems. We provide an overview of transmit beamforming designs for various energy efficiency (EE) metrics including maximizing the overall network EE, sum weighted EE, and fairness EE. Generally, an EE optimization problem is a nonconvex program for which finding the globally optimal solutions requires high computational effort. Consequently, several low-complexity suboptimal approaches have been proposed. Here, we sum up the main concepts of the recently proposed algorithms based on the state-of-the-art successive convex approximation (SCA) framework. Moreover, we discuss the application to the newly posted EE problems including new EE metrics and power consumption models. Furthermore, distributed implementation developed based on alternating direction method of multipliers (ADMM) for the provided solutions is also discussed. For the sake of completeness, we provide numerical comparison of the SCA based approaches and the conventional solutions developed based on parametric transformations (PTs). We also demonstrate the differences and roles of different EE objectives and power consumption models.
Abstract Energy efficiency (EE) is becoming one of the important criteria in wireless transmission design. This paper discusses the recently proposed energy-efficient transmit beamforming designs for multicell multiuser multiple-input single-output (MISO) systems, including maximizing overall network EE, sum weighted EE and fairness EE. Generally, the EE optimization problems are NP-hard nonconvex programs for which finding the globally optimal solutions is challenging. For low-complexity suboptimal approaches, there is a class of solutions conventionally developed based on parametric transformations. However, those have been revealed problematic in terms of computational complexity and convergence. To overcome these issues, novel algorithms have been recently developed based on the state-of-the-art successive convex approximation (SCA) framework. Here we sum up the basic concepts of the algorithms and provide numerical results which illustrate the solution quality compared to the existing methods.
Abstract We investigate the fairness of achievable energy efficiency in a multicell multiuser multiple-input single-output (MISO) downlink system, where a beamforming scheme is designed to maximize the minimum energy efficiency among all base stations. The resulting optimization problem is a nonconvex max-min fractional program, which is generally difficult to solve optimally. We propose an iterative beamformer design based on an inner approximation algorithm which aims at locating a Karush-Kuhn-Tucker solution to the nonconvex program. By novel transformations, we arrive at a convex problem at each iteration of the proposed algorithm, which is amendable for being approximated by a second order cone program. The numerical results demonstrate that the proposed algorithm outperforms the existing schemes in terms of the convergence rate and processing time.