Analog RF Computing: A New Paradigm for Energy-Efficient Edge AI Over MU-MIMO Systems

2026-05-13

AI Summary

This work addresses the high memory footprint and energy consumption of conventional digital computing in edge AI inference by proposing an analog radio-frequency (RF) computing framework tailored for MU-MIMO systems. It introduces, for the first time, a computation-oriented (rather than communication-oriented) physical-layer architecture wherein base stations broadcast RF waveforms encoding neural network weights, and client devices perform matrix-vector multiplication directly via passive mixers, enabling ultra-low-power inference. The framework supports fine-grained precision control at both client and network levels and jointly optimizes uniform and mixed-precision inference through a synergistic design incorporating analytically tractable modeling, base-station beamforming, client-side scaling, and a low-complexity non-convex optimization algorithm. Simulations under 3GPP standards demonstrate nearly two orders of magnitude reduction in client energy consumption compared to digital baselines, with mixed-precision strategies further enhancing energy efficiency.

Abstract

Modern edge devices increasingly rely on neural networks for intelligent applications. However, conventional digital computing-based edge inference requires substantial memory and energy consumption. In analog radio frequency (RF) computing, a base station (BS) encodes the weights of the neural networks and broadcasts the RF waveforms to the clients. Each client reuses its passive mixer to multiply the received weight-encoded waveform with a locally generated input-encoded waveform. This enables wireless receivers to perform the matrix-vector multiplications (MVMs) that account for most of the computation burden in edge inference with ultra-low energy consumption. Unlike conventional downlink transmissions which are optimized for communications, analog RF computing requires a computing-centric physical layer that controls both the analog MVM accuracy and the energy consumption for inference. Motivated by this, in this paper, we propose a physical layer design framework for analog RF computing in MU-MIMO wireless systems. We derive tractable models for computing accuracy and energy consumption for inference, formulate a joint BS beamforming and client-side scaling problem subject to computing accuracy, transmit power, and hardware constraints, and develop a low-complexity algorithm to solve the non-convex problem. The proposed design provides client- and layer-specific accuracy control for both uniform- and mixed-precision inference. Simulations under 3GPP specifications show that analog RF computing can significantly reduce client-side energy consumption by nearly two orders of magnitude compared to digital computing, while mixed-precision inference requires even lower energy consumption than uniform-precision inference. Overall, these results establish analog RF computing over wireless networks as a promising paradigm for energy-efficient edge inference.

Problem

• Analog RF Computing
• Edge AI
• Energy Efficiency
• MU-MIMO
• Matrix-Vector Multiplication

Innovation

• Analog RF Computing
• MU-MIMO
• Energy-Efficient Edge AI
• Matrix-Vector Multiplication
• Mixed-Precision Inference