Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Near-optimal stochastic MIMO signal detection with a mixture of t-distribution prior

Junichiro Hagiwara, Kazushi Matsumura, Hiroki Asumi, Yukiko Kasuga, Toshihiko Nishimura, Takanori Sato, Yasutaka Ogawa, Takeo Ohgane

TL;DR

This paper tackles the challenge of robust MIMO signal detection under practical complexity by reframing the discrete prior problem as a continuous one and solving the posterior with Hamiltonian Monte Carlo. It introduces a mixture of $t$-distributions as the prior, which sharpens focus around potential signal points while retaining exploration in the surrounding space, enabling near-optimal detection with polynomial complexity. Empirical results show substantial BER gains over traditional discrete-prior baselines, particularly for higher-order modulations, and demonstrate performance approaching that of a SISO AWGN benchmark in favorable regimes. The approach offers practical implications for future wireless systems and suggests a generalizable strategy for applying continuous priors to combinatorial inference problems in communication and beyond.

Abstract

Multiple-input multiple-output (MIMO) systems will play a crucial role in future wireless communication, but improving their signal detection performance to increase transmission efficiency remains a challenge. To address this issue, we propose extending the discrete signal detection problem in MIMO systems to a continuous one and applying the Hamiltonian Monte Carlo method, an efficient Markov chain Monte Carlo algorithm. In our previous studies, we have used a mixture of normal distributions for the prior distribution. In this study, we propose using a mixture of t-distributions, which further improves detection performance. Based on our theoretical analysis and computer simulations, the proposed method can achieve near-optimal signal detection with polynomial computational complexity. This high-performance and practical MIMO signal detection could contribute to the development of the 6th-generation mobile network.

Near-optimal stochastic MIMO signal detection with a mixture of t-distribution prior

TL;DR

This paper tackles the challenge of robust MIMO signal detection under practical complexity by reframing the discrete prior problem as a continuous one and solving the posterior with Hamiltonian Monte Carlo. It introduces a mixture of -distributions as the prior, which sharpens focus around potential signal points while retaining exploration in the surrounding space, enabling near-optimal detection with polynomial complexity. Empirical results show substantial BER gains over traditional discrete-prior baselines, particularly for higher-order modulations, and demonstrate performance approaching that of a SISO AWGN benchmark in favorable regimes. The approach offers practical implications for future wireless systems and suggests a generalizable strategy for applying continuous priors to combinatorial inference problems in communication and beyond.

Abstract

Multiple-input multiple-output (MIMO) systems will play a crucial role in future wireless communication, but improving their signal detection performance to increase transmission efficiency remains a challenge. To address this issue, we propose extending the discrete signal detection problem in MIMO systems to a continuous one and applying the Hamiltonian Monte Carlo method, an efficient Markov chain Monte Carlo algorithm. In our previous studies, we have used a mixture of normal distributions for the prior distribution. In this study, we propose using a mixture of t-distributions, which further improves detection performance. Based on our theoretical analysis and computer simulations, the proposed method can achieve near-optimal signal detection with polynomial computational complexity. This high-performance and practical MIMO signal detection could contribute to the development of the 6th-generation mobile network.
Paper Structure (19 sections, 8 equations, 6 figures, 2 tables, 1 algorithm)

This paper contains 19 sections, 8 equations, 6 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Problem formulation
  3. System model
  4. Likelihood
  5. Prior distribution
  6. Posterior distribution
  7. Previous work
  8. MGS method MGS
  9. EP method EP
  10. Stochastic signal detection with a mixture of $t$-distributions prior
  11. Setting the prior distribution
  12. Approximation algorithm for the posterior distribution
  13. Numerical results and discussion
  14. Settings in numerical analysis
  15. Computational complexity
  16. ...and 4 more sections

Figures (6)

  • Figure 1: Two examples of priors for BPSK.
  • Figure 2: The normal and $t$-distributions with the same scale parameter $\sigma$.
  • Figure 3: Laplace (double exponential) distribution.
  • Figure 4: Average BER vs. average received SNR for $\rho = 0 \text{ and } 0.5$.
  • Figure 5: Average BER vs. average received SNR for $\rho = 0$.
  • ...and 1 more figures