Geo2ComMap: Deep Learning-Based MIMO Throughput Prediction Using Geographic Data

2025-04-01

AI Summary

This work addresses the high computational complexity and overhead associated with full-coverage throughput (Tput) prediction in MIMO systems. We propose a geospatial information–driven end-to-end deep learning framework that jointly leverages CNN and Transformer architectures, incorporates explicit geospatial feature encoding, and employs a sparse sampling strategy to suppress outlier prediction errors. Crucially, the method bypasses explicit channel matrix modeling and enables joint regression of multiple performance metrics—including throughput—directly from geographic and system-level inputs. Our key contributions are: (i) the first establishment of a geographically aware MIMO performance mapping paradigm; and (ii) accurate end-to-end throughput prediction across the full 0–1900 Mbps range, achieving a median absolute error of only 27.35 Mbps. Compared to conventional model-based or simulation-driven approaches, our method significantly reduces computational overhead and inference latency while maintaining high fidelity.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate communication performance prediction is crucial for wireless applications such as network deployment and resource management. Unlike conventional systems with a single transmit and receive antenna, throughput (Tput) estimation in antenna array-based multiple-output multiple-input (MIMO) systems is computationally intensive, i.e., requiring analysis of channel matrices, rank conditions, and spatial channel quality. These calculations impose significant computational and time burdens. This paper introduces Geo2ComMap, a deep learning-based framework that leverages geographic databases to efficiently estimate multiple communication metrics across an entire area in MIMO systems using only sparse measurements. To mitigate extreme prediction errors, we propose a sparse sampling strategy. Extensive evaluations demonstrate that Geo2ComMap accurately predicts full-area communication metrics, achieving a median absolute error of 27.35 Mbps for Tput values ranging from 0 to 1900 Mbps.

Problem

• Predict MIMO throughput using geographic data efficiently
• Reduce computational burden in MIMO performance estimation
• Minimize extreme errors in communication metrics prediction

Innovation

• Deep learning predicts MIMO throughput geographically
• Sparse sampling reduces extreme prediction errors
• Leverages geographic data for efficient metric estimation