Detection by Sampling: Massive MIMO Detector based on Langevin Dynamics

Zilberstein, Nicolas; Dick, Chris; Doost-Mohammady, Rahman; Sabharwal, Ashutosh; Segarra, Santiago

doi:10.48550/arxiv.2202.12199

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Notice that this method can be described by (6) by adding a randomized step to generate random samples. Lastly, a detector using the Langevin dynamics was independently proposed in [28] (concurrently made available with our preliminary work [1]), but the approach is significantly different from the one presented in this paper, as they do not adopt an annealing process to incorporate the discrete nature of the problem, they are constrained to a particular modulation scheme, and they do not propose a robust learning-based version rooted in algorithm unfolding.…”

Section: B Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Annealed Langevin Dynamics for Massive MIMO Detection

Zilberstein¹,

Dick²,

Doost-Mohammady³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Solving the optimal symbol detection problem in multiple-input multiple-output (MIMO) systems is known to be NP-hard. Hence, the objective of any detector of practical relevance is to get reasonably close to the optimal solution while keeping the computational complexity in check. In this work, we propose a MIMO detector based on an annealed version of Langevin (stochastic) dynamics. More precisely, we define a stochastic dynamical process whose stationary distribution coincides with the posterior distribution of the symbols given our observations. In essence, this allows us to approximate the maximum a posteriori estimator of the transmitted symbols by sampling from the proposed Langevin dynamic. Furthermore, we carefully craft this stochastic dynamic by gradually adding a sequence of noise with decreasing variance to the trajectories, which ensures that the estimated symbols belong to a prespecified discrete constellation. Based on the proposed MIMO detector, we also design a robust version of the method by unfolding and parameterizing one term -the score of the likelihood -by a neural network. Through numerical experiments in both synthetic and real-world data, we show that our proposed detector yields state-of-the-art symbol error rate performance and the robust version becomes noise-variance agnostic.

show abstract

Section: B Related Workmentioning

confidence: 99%

“…Email: {nzilberstein, doost, ashu, segarra}@rice.edu, cdick@nvidia.com. Preliminary results were published in[1].…”

mentioning

confidence: 99%