Robust Resource Allocation for Multi-Antenna URLLC-OFDMA Systems in a Smart Factory

TL;DR

This paper develops a robust beamforming and resource allocation method for multi-antenna URLLC-OFDMA systems in smart factories, optimizing power while ensuring ultra-low latency and high reliability.

Contribution

It introduces a novel non-convex penalty approach for sparse resource block assignment and demonstrates its advantages over existing methods in power efficiency and robustness.

Findings

NCP approach outperforms reweighted ℓ₁ in power and convergence.

System performance improves with more antennas and lower latency.

Robustness to channel uncertainty is enhanced by the proposed method.

Abstract

In this paper, we investigate the worst-case robust beamforming design and resource block (RB) assignment problem for total transmit power minimization of the central controller while guaranteeing each robot's transmission with target number of data bits and within required ultra-low latency and extremely high reliability. By using the property of the independence of each robot's beamformer design, we can obtain the equivalent power control design form of the original beamforming design. The binary RB mapping indicators are transformed into continuous ones with additional $\ell_0$-norm constraints to promote sparsity on each RB. A novel non-convex penalty (NCP) approach is applied to solve such $\ell_0$-norm constraints. Numerical results demonstrate the superiority of the NCP approach to the well-known reweighted $\ell_1$ method in terms of the optimized power consumption, convergence rate and robustness to channel realizations. Also, the impacts of latency, reliability, number of transmit antennas and channel uncertainty on the system performance are revealed.