Abstract:Reconfigurable intelligent surface (RIS) and hybrid beamforming have been envisioned as promising alternatives to alleviate blockage vulnerability and enhance coverage capability for terahertz (THz) multi-user multiple-input multiple-output systems that suffer from severe propagation attenuation and poor diffraction. Considering that the joint beamforming with large-scale array elements at transceivers and RIS is extremely complicated, the codebook based beamforming can be employed in a computationally efficient manner. However, the codeword selection for analog beamforming is an intractable combinatorial optimization (CO) problem. To this end, an iterative alternating search (IAS) algorithm is developed to achieve a near-optimal sum-rate performance with low computational complexity in contrast with the optimal exhaustive search algorithm. According to the THz channel dataset generated by the IAS algorithm, a multi-task learning based analog beam selection (MTL-ABS) framework is developed to further decrease the computation overhead. Specifically, we take the CO problem as a multi-task classification problem and implement multiple beam selection tasks at transceivers and RIS simultaneously. Remarkably, residual network and self-attention mechanism are used to combat the network degradation and mine intrinsic THz channel features. Finally, blockwise convergence analysis and numerical results demonstrate the effectiveness of the MTL-ABS framework over search based counterparts.

Stabilizing Coherently Combined Beam Power Using a Robust Learning Algorithm

Experimental Beam Combining Stabilization Using Machine Learning Trained While Phases Drift

Stabilization of the 81-channel coherent beam combination using machine learning

Deterministic Stabilization of Eight-Way 2D Diffractive Beam Combining Using Pattern Recognition.

Method for enhancing stability in multi-beam microscopy

Artificial Neural Networks Applied to Stabilization of 81-Beam Coherent Combining

Method for Active Spatial Alignment and Stabilization of Laser Beams in Multi-Beam Systems

Diffractive combining and control of femtosecond pulse beam arrays

Characterization and Control of 81-Beam Diffractive Coherent Combining

High environmentally adaptable phase control with reinforcement learning for coherent beam combination

Single-step phase identification and phase locking for coherent beam combination using deep learning

Machine Learning Pattern Recognition Algorithm with Applications to Coherent Laser Combination

81-Beam Coherent Combination Using a Programmable Array Generator.

Deep learning-based phase control method for coherent beam combining and its application in generating orbital angular momentum beams

A phase-error prediction method for coherent beam combining via convolutional neural network

Scalable structure of coherent polarization beam combining based on tapered diode laser amplifiers

CALIPR: Coherent Addition using Learned Interference Pattern Recognition

Injection Locking Method for Raman Beams in Atom Interferometer

Deep Learning of Coherent Laser Arrays in Angular Domain for Orbital Angular Momentum Beams Customization

Beam Selection for RIS-Enabled Terahertz Multi-User MIMO Systems via Multi-Task Learning

Coherent beam combination of adaptive fiber laser array with tilt-tip and phase-locking control