Study on precoding optimization algorithms in massive MIMO system with multi-antenna users

Abstract

The paper studies the multi-user precoding problem as a non-convex optimization problem for wireless multiple inputs and multiple outputs (MIMO) systems. In our work, we approximate the target Spectral Efficiency function with a novel computationally simpler function. Then, we reduce the precoding problem to an unconstrained optimization task using a special differential projection method and solve it by the Quasi-Newton L-BFGS iterative procedure to achieve gains in capacity. We are testing the proposed approach in several scenarios generated using Quadriga-open-source software for generating realistic radio channel impulse response. Our method shows monotonic improvement over heuristic methods with reasonable computation time. The proposed L-BFGS optimization scheme is novel in this area and shows a significant advantage over the standard approaches. The proposed method has simple implementation and can be a good reference for other heuristic algorithms in this field.

Keywords:

• Optimization
• massive MIMO
• precoding
• SVD
• L-BFGS
• interior-point

Acknowledgments

Authors are grateful for D. Minenkov for the fruitful discussions.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by Huawei Technologies.

Notes on contributors

Evgeny Bobrov

Evgeny Bobrov is a PhD student of the Department of Computational Mathematics and Cybernetics, Lomonosov Moscow State University, Russian Federation. Graduated from Lomonosov Moscow State University with Bachelor's and Master's degrees in the same university. The topic of my scientific qualification work is: “Machine Learning and Optimization Algorithms in Wireless Communications Problems”. Experience of teaching 4 years. Area of interest includes machine learning, optimization and modern wireless communication problems.

Dmitry Kropotov

Dmitry Kropotov is a specialist: Lomonosov Moscow State University. M.V. Lomonosov Moscow State University, specialty “Applied mathematics and informatics”, qualification “mathematician, system programmer”. Professional interests: optimization, Bayesian methods, neural network models, applied data analysis. Work experience: February 2007 - December 2013, Junior Researcher, A.A. Dorodnitsyn Computer Center RAS. February 2014 - present, researcher, Lomonosov Moscow State University, Department of Computational Mathematics and Cybernetics, Department of Mathematical Methods of Forecasting. June 2014-present, Scientific Secretary, Lomonosov Moscow State University, Department of Computational Mathematics and Cybernetics, Department of Mathematical Methods of Forecasting. February 2018 - present, Senior Researcher, Higher School of Economics-Samsung Joint Laboratory.

Sergey Troshin

Sergey Troshin is a Research Intern: Faculty of Computer Science / Department of Big Data and Information Retrieval / Center for Deep Learning and Bayesian Methods. Started working at NRU HSE in 2017. Research and teaching experience: 1 year, 6 months. Responsibilities: to conduct research for Centre of Deep Learning and Bayesian Methods. Professional Interests: Implicit models, automatic source code analysis. Employment history: Research Assistant at Advanced Research NLP Group ABBYY, June-August 2019. Development of a system for active learning of neural networks to search for relationships between entities in a text.

Danila Zaev

Danila Zaev received Ph.D degree in mathematics in 2017 at National Research University Higher School of Economics (SU HSE), Department of Mathematics. Title of the PhD thesis: Monge-Kantorovich problem with linear constraints. His professional Interests: wireless communication, massive MIMO, stochastic processes, mathematical optimization.