-
Artificial Intelligence-Defined 5G Radio Access Networks
Authors:
Miao Yao,
Munawwar Sohul,
Vuk Marojevic,
Jeffrey H. Reed
Abstract:
Massive multiple-input multiple-output antenna systems, millimeter wave communications, and ultra-dense networks have been widely perceived as the three key enablers that facilitate the development and deployment of 5G systems. This article discusses the intelligent agent in 5G base station which combines sensing, learning, understanding and optimizing to facilitate these enablers. We present a fl…
▽ More
Massive multiple-input multiple-output antenna systems, millimeter wave communications, and ultra-dense networks have been widely perceived as the three key enablers that facilitate the development and deployment of 5G systems. This article discusses the intelligent agent in 5G base station which combines sensing, learning, understanding and optimizing to facilitate these enablers. We present a flexible, rapidly deployable, and cross-layer artificial intelligence (AI)-based framework to enable the imminent and future demands on 5G and beyond infrastructure. We present example AI-enabled 5G use cases that accommodate important 5G-specific capabilities and discuss the value of AI for enabling beyond 5G network evolution.
△ Less
Submitted 22 December, 2018; v1 submitted 21 November, 2018;
originally announced November 2018.
-
Semidefinite Relaxation-Based PAPR-Aware Precoding for Massive MIMO-OFDM Systems
Authors:
Miao Yao,
Matt Carrick,
Munawwar M. Sohul,
Vuk Marojevic,
Cameron D. Patterson,
Jeffrey H. Reed
Abstract:
Massive MIMO requires a large number of antennas and the same amount of power amplifiers (PAs), one per antenna. As opposed to 4G base stations, which could afford highly linear PAs, next-generation base stations will need to use inexpensive PAs, which have a limited region of linear amplification. One of the research challenges is effectively handling signals which have high peak-to-average power…
▽ More
Massive MIMO requires a large number of antennas and the same amount of power amplifiers (PAs), one per antenna. As opposed to 4G base stations, which could afford highly linear PAs, next-generation base stations will need to use inexpensive PAs, which have a limited region of linear amplification. One of the research challenges is effectively handling signals which have high peak-to-average power ratios (PAPRs), such as orthogonal frequency division multiplexing (OFDM). This paper introduces a PAPR-aware precoding scheme that exploits the excessive spatial degrees-of-freedom of large scale multiple-input multipleoutput (MIMO) antenna systems. This typically requires finding a solution to a nonconvex optimization problem. Instead of relaxing the problem to minimize the peak power, we introduce a practical semidefinite relaxation (SDR) framework that enables accurately and efficiently approximating the theoretical PAPR-aware precoding performance for OFDM-based massive MIMO systems. The framework allows incorporating channel uncertainties and intercell coordination. Numerical results show that several orders of magnitude improvements can be achieved w.r.t. state of the art techniques, such as instantaneous power consumption reduction and multiuser interference cancellation. The proposed PAPRaware precoding can be effectively handled along with the multicell signal processing by the centralized baseband processing platforms of next-generation radio access networks. Performance can be traded for the computing efficiency for other platforms
△ Less
Submitted 2 October, 2018;
originally announced October 2018.
-
A Digital Predistortion Scheme Exploiting Degrees-of-Freedom for Massive MIMO Systems
Authors:
Miao Yao,
Munawwar Sohul,
Randall Nealy,
Vuk Marojevic,
Jeffrey Reed
Abstract:
The primary source of nonlinear distortion in wireless transmitters is the power amplifier (PA). Conventional digital predistortion (DPD) schemes use high-order polynomials to accurately approximate and compensate for the nonlinearity of the PA. This is not practical for scaling to tens or hundreds of PAs in massive multiple-input multiple-output (MIMO) systems. There is more than one candidate pr…
▽ More
The primary source of nonlinear distortion in wireless transmitters is the power amplifier (PA). Conventional digital predistortion (DPD) schemes use high-order polynomials to accurately approximate and compensate for the nonlinearity of the PA. This is not practical for scaling to tens or hundreds of PAs in massive multiple-input multiple-output (MIMO) systems. There is more than one candidate precoding matrix in a massive MIMO system because of the excess degrees-of-freedom (DoFs), and each precoding matrix requires a different DPD polynomial order to compensate for the PA nonlinearity. This paper proposes a low-order DPD method achieved by exploiting massive DoFs of next-generation front ends. We propose a novel indirect learning structure which adapts the channel and PA distortion iteratively by cascading adaptive zero forcing precoding and DPD. Our solution uses a 3rd order polynomial to achieve the same performance as the conventional DPD using an 11th order polynomial for a 100x10 massive MIMO configuration. Experimental results show a 70% reduction in computational complexity, enabling ultra-low latency communications.
△ Less
Submitted 18 January, 2018;
originally announced January 2018.
-
Sustainable Green Networking: Exploiting Degrees of Freedom towards Energy-Efficient 5G Systems
Authors:
Miao Yao,
Munawwar Sohul,
Xiaofu Ma,
Vuk Marojevic,
Jeffrey H. Reed
Abstract:
The carbon footprint concern in the development and deployment of 5G new radio systems has drawn the attention to several stakeholders. In this article, we analyze the critical power consuming component of all candidate 5G system architectures-the power amplifier (PA)-and propose PA-centric resource management solutions for green 5G communications. We discuss the impact of ongoing trends in cellul…
▽ More
The carbon footprint concern in the development and deployment of 5G new radio systems has drawn the attention to several stakeholders. In this article, we analyze the critical power consuming component of all candidate 5G system architectures-the power amplifier (PA)-and propose PA-centric resource management solutions for green 5G communications. We discuss the impact of ongoing trends in cellular communications on sustainable green networking and analyze two communications architectures that allow exploiting the extra degrees-of-freedom (DoF) from multi-antenna and massive antenna deployments: small cells/distributed antenna network and massive MIMO. For small cell systems with a moderate number of antennas, we propose a peak to average power ratio-aware resource allocation scheme for joint orthogonal frequency and space division multiple access. For massive MIMO systems, we develop a highly parallel recurrent neural network for energy-efficient precoding. Simulation results for representative 5G deployment scenarios demonstrate an energy efficiency improvement of one order of magnitude or higher with respect to current state-of-the-art solutions.
△ Less
Submitted 1 November, 2017;
originally announced November 2017.
