Abstract:In an aerial hybrid massive multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) system, how to design a spectral-efficient broadband multi-user hybrid beamforming with a limited pilot and feedback overhead is challenging. To this end, by modeling the key transmission modules as an end-to-end (E2E) neural network, this paper proposes a data-driven deep learning (DL)-based unified hybrid beamforming framework for both the time division duplex (TDD) and frequency division duplex (FDD) systems with implicit channel state information (CSI). For TDD systems, the proposed DL-based approach jointly models the uplink pilot combining and downlink hybrid beamforming modules as an E2E neural network. While for FDD systems, we jointly model the downlink pilot transmission, uplink CSI feedback, and downlink hybrid beamforming modules as an E2E neural network. Different from conventional approaches separately processing different modules, the proposed solution simultaneously optimizes all modules with the sum rate as the optimization object. Therefore, by perceiving the inherent property of air-to-ground massive MIMO-OFDM channel samples, the DL-based E2E neural network can establish the mapping function from the channel to the beamformer, so that the explicit channel reconstruction can be avoided with reduced pilot and feedback overhead. Besides, practical low-resolution phase shifters (PSs) introduce the quantization constraint, leading to the intractable gradient backpropagation when training the neural network. To mitigate the performance loss caused by the phase quantization error, we adopt the transfer learning strategy to further fine-tune the E2E neural network based on a pre-trained network that assumes the ideal infinite-resolution PSs. Numerical results show that our DL-based schemes have considerable advantages over state-of-the-art schemes.

Beamforming Design for Large-Scale Antenna Arrays Using Deep Learning

Design of Hybrid Beamforming for Multiuser MIMO mmWave Systems Using Deep Learning

Sparse-reconstruction-based high resolution beamforming and its application to multi-beam systems

Deep Learning Based Joint Beam Selection and Precoding Design for Mmwave Systems with Lens Arrays

Robust mmWave Beamforming by Self-Supervised Hybrid Deep Learning

Near-field Beam training for Extremely Large-scale MIMO Based on Deep Learning

Sparse Array Design for Optimum Beamforming Using Deep Learning

Deep Learning Based Hybrid Beamforming for mmWave Dual-Functional Radar-Communication

Deep Learning of Near Field Beam Focusing in Terahertz Wideband Massive MIMO Systems

Deep Learning based Fast and Accurate Beamforming for Millimeter-Wave Systems

Sparse Array Capon Beamformer Design Availing Deep Learning

Mixed-Timescale Deep-Unfolding for Joint Channel Estimation and Hybrid Beamforming

Radio Frequency Beamforming Based on a Complex Domain Frontend

Hierarchical Beamforming for Broadcast Channels

Deep Learning for Beam-Management: State-of-the-Art, Opportunities and Challenges

Deep Learning for Joint Design of Pilot, Channel Feedback, and Hybrid Beamforming in FDD Massive MIMO-OFDM Systems

Deep Learning for Mmwave Beam-Management: State-of-the-Art, Opportunities and Challenges

Reinforcement Learning for Beam Pattern Design in Millimeter Wave and Massive MIMO Systems

Data-Driven Deep Learning Based Hybrid Beamforming for Aerial Massive MIMO-OFDM Systems with Implicit CSI

Beamfocusing Optimization for Near-Field Wideband Multi-User Communications

Technique for Large-Scale Antenna Beamforming Based on Neural Network