A Low-complexity Structured Neural Network Approach to Intelligently Realize Wideband Multi-beam Beamformers

2025-03-26

AI Summary

To address the high spatial and computational complexity hindering real-time implementation of broadband multi-beam beamformers, this paper proposes a structured sparse neural network architecture. The method innovatively embeds the inherent structural prior of the delay Vandermonde matrix (DVM) into the network weight design, employing submatrix-level sparsity constraints to model wideband antenna array responses—thereby enabling dispersion-free, concurrent multi-beam formation with low latency. Theoretical analysis shows the computational complexity is reduced to *O*(*pLM* log *M*), substantially outperforming conventional fully connected networks with *O*(*M*²*L*). Simulation results over the 24–32 GHz band demonstrate that the proposed approach achieves mean squared error performance comparable to fully connected networks while significantly reducing memory footprint and computational overhead. This enables energy-efficient, real-time broadband multi-beamforming for intelligent wireless systems.

Technology Category

• Machine Learning: Matrix & Tensor Methods
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

True-time-delay (TTD) beamformers can produce wideband, squint-free beams in both analog and digital signal domains, unlike frequency-dependent FFT beams. Our previous work showed that TTD beamformers can be efficiently realized using the elements of delay Vandermonde matrix (DVM), answering the longstanding beam-squint problem. Thus, building on our work on classical algorithms based on DVM, we propose neural network (NN) architecture to realize wideband multi-beam beamformers using structure-imposed weight matrices and submatrices. The structure and sparsity of the weight matrices and submatrices are shown to reduce the space and computational complexities of the NN greatly. The proposed network architecture has O(pLM logM) complexity compared to a conventional fully connected L-layers network with O(M2L) complexity, where M is the number of nodes in each layer of the network, p is the number of submatrices per layer, and M>>p. We will show numerical simulations in the 24 GHz to 32 GHz range to demonstrate the numerical feasibility of realizing wideband multi-beam beamformers using the proposed neural architecture. We also show the complexity reduction of the proposed NN and compare that with fully connected NNs, to show the efficiency of the proposed architecture without sacrificing accuracy. The accuracy of the proposed NN architecture was shown using the mean squared error, which is based on an objective function of the weight matrices and beamformed signals of antenna arrays, while also normalizing nodes. The proposed NN architecture shows a low-complexity NN realizing wideband multi-beam beamformers in real-time for low-complexity intelligent systems.

Problem

• Develop low-complexity neural network for wideband multi-beam beamforming
• Address beam-squint issue using structured weight matrices
• Reduce computational complexity without sacrificing accuracy

Innovation

• Neural network with structured weight matrices
• Low-complexity O(pLM logM) architecture
• Wideband multi-beam beamforming solution