-
Machine Learning Methods for User Positioning With Uplink RSS in Distributed Massive MIMO
Authors:
K. N. R. Surya Vara Prasad,
Ekram Hossain,
Vijay K. Bhargava
Abstract:
We consider a machine learning approach based on Gaussian process regression (GP) to position users in a distributed massive multiple-input multiple-output (MIMO) system with the uplink received signal strength (RSS) data. We focus on the scenario where noise-free RSS is available for training, but only noisy RSS is available for testing purposes. To estimate the test user locations and their 2σ e…
▽ More
We consider a machine learning approach based on Gaussian process regression (GP) to position users in a distributed massive multiple-input multiple-output (MIMO) system with the uplink received signal strength (RSS) data. We focus on the scenario where noise-free RSS is available for training, but only noisy RSS is available for testing purposes. To estimate the test user locations and their 2σ error-bars, we adopt two state-of-the-art GP methods, namely, the conventional GP (CGP) and the numerical approximation GP (NaGP) methods. We find that the CGP method, which treats the noisy test RSS vectors as noise-free, provides unrealistically small 2σ error-bars on the estimated locations. To alleviate this concern, we derive the true predictive distribution for the test user locations and then employ the NaGP method to numerically approximate it as a Gaussian with the same first and second order moments. We also derive a Bayesian Cramer-Rao lower bound (BCRLB) on the achievable root- mean-squared-error (RMSE) performance of the two GP methods. Simulation studies reveal that: (i) the NaGP method indeed provides realistic 2σ error-bars on the estimated locations, (ii) operation in massive MIMO regime improves the RMSE performance, and (iii) the achieved RMSE performances are very close to the derived BCRLB.
△ Less
Submitted 19 January, 2018;
originally announced January 2018.
-
Low-Dimensionality of Noise-Free RSS and its Application in Distributed Massive MIMO
Authors:
K. N. R. Surya Vara Prasad,
Ekram Hossain,
Vijay K. Bhargava
Abstract:
We examine the dimensionality of noise-free uplink received signal strength (RSS) data in a distributed multiuser massive multiple-input multiple-output system. Specifically, we apply principal component analysis to the noise-free uplink RSS and observe that it has a low-dimensional principal subspace. We make use of this unique property to propose RecGP - a reconstruction-based Gaussian process r…
▽ More
We examine the dimensionality of noise-free uplink received signal strength (RSS) data in a distributed multiuser massive multiple-input multiple-output system. Specifically, we apply principal component analysis to the noise-free uplink RSS and observe that it has a low-dimensional principal subspace. We make use of this unique property to propose RecGP - a reconstruction-based Gaussian process regression (GP) method which predicts user locations from uplink RSS data. Considering noise-free RSS for training and noisy test RSS for location prediction, RecGP reconstructs the noisy test RSS from a low- dimensional principal subspace of the noise-free training RSS. The reconstructed RSS is input to a trained GP model for location prediction. Noise reduction facilitated by the reconstruction step allows RecGP to achieve lower prediction error than standard GP methods which directly use the test RSS for location prediction.
△ Less
Submitted 7 August, 2017;
originally announced August 2017.
-
Energy Efficiency in Massive MIMO-Based 5G Networks: Opportunities and Challenges
Authors:
K. N. R. Surya Vara Prasad,
Ekram Hossain,
Vijay K. Bhargava
Abstract:
As we make progress towards the era of fifth generation (5G) communication networks, energy efficiency (EE) becomes an important design criterion because it guarantees sustainable evolution. In this regard, the massive multiple-input multiple-output (MIMO) technology, where the base stations (BSs) are equipped with a large number of antennas so as to achieve multiple orders of spectral and energy…
▽ More
As we make progress towards the era of fifth generation (5G) communication networks, energy efficiency (EE) becomes an important design criterion because it guarantees sustainable evolution. In this regard, the massive multiple-input multiple-output (MIMO) technology, where the base stations (BSs) are equipped with a large number of antennas so as to achieve multiple orders of spectral and energy efficiency gains, will be a key technology enabler for 5G. In this article, we present a comprehensive discussion on state-of-the-art techniques which further enhance the EE gains offered by massive MIMO (MM). We begin with an overview of MM systems and discuss how realistic power consumption models can be developed for these systems. Thereby, we discuss and identify few shortcomings of some of the most prominent EE-maximization techniques present in the current literature. Then, we discuss "hybrid MM systems" operating in a 5G architecture, where MM operates in conjunction with other potential technology enablers, such as millimetre wave, heterogenous networks, and energy harvesting networks. Multiple opportunities and challenges arise in such a 5G architecture because these technologies benefit mutually from each other and their coexistence introduces several new constraints on the design of energy-efficient systems. Despite clear evidence that hybrid MM systems can achieve significantly higher EE gains than conventional MM systems, several open research problems continue to roadblock system designers from fully harnessing the EE gains offered by hybrid MM systems. Our discussions lead to the conclusion that hybrid MM systems offer a sustainable evolution towards 5G networks and are therefore an important research topic for future work.
△ Less
Submitted 27 November, 2015;
originally announced November 2015.
