Programmable and GPU-Accelerated Edge Inference for Real-Time ISAC on NVIDIA ARC-OTA

2025-12-06

AI Summary

To address 6G integrated sensing and communication (ISAC) requirements, existing cellular networks struggle to simultaneously achieve high-precision sensing and communication performance under limited bandwidth, while relying on dedicated hardware and protocol stack modifications. This paper proposes the first programmable edge AI inference framework for ISAC, built upon an Open RAN decentralized application (dApp) architecture and accelerated by GPU—deployed on the NVIDIA ARC-OTA platform—to enable real-time 5G NR physical-layer signal processing and fully software-defined ISAC functionality. The framework integrates dynamic multipath cancellation, a lightweight neural network, and CUDA-optimized inference, requiring no RAN protocol stack changes or specialized hardware. It achieves sub-millisecond channel state information (CSI) extraction and high-accuracy indoor localization. Experimental evaluation on a 3GPP-compliant 5G system demonstrates an average positioning error of 77 cm, with 75% of results under 1 m and end-to-end latency below 0.5 ms.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Learning on the Edge & Model Compression
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

The transition of cellular networks to (i) software-based systems on commodity hardware and (ii) platforms for services beyond connectivity introduces critical system-level challenges. As sensing emerges as a key feature toward 6G standardization, supporting Integrated Sensing and Communication (ISAC) with limited bandwidth and piggybacking on communication signals, while maintaining high reliability and performance, remains a fundamental challenge. In this paper, we provide two key contributions. First, we present a programmable, plug-and-play framework for processing PHY/MAC signals through real-time, GPU-accelerated Artificial Intelligence (AI) applications on the edge Radio Access Network (RAN) infrastructure. Building on the Open RAN dApp architecture, the framework interfaces with a GPU-accelerated gNB based on NVIDIA ARC-OTA, feeding PHY/MAC data to custom AI logic with latency under 0.5 ms for complex channel state information extraction. Second, we demonstrate the framework's capabilities through cuSense, an indoor localization dApp that consumes uplink DMRS channel estimates, removes static multipath components, and runs a neural network to infer the position of a moving person. Evaluated on a 3GPP-compliant 5G NR deployment, cuSense achieves a mean localization error of 77 cm, with 75% of predictions falling within 1 meter. This is without dedicated sensing hardware or modifications to the RAN stack or signals. We plan to release both the framework and cuSense pipelines as open source, providing a reference design for future AI-native RANs and ISAC applications.

Problem

• Enabling real-time ISAC with limited bandwidth and high reliability
• Providing programmable edge AI for PHY/MAC signal processing in RAN
• Achieving accurate indoor localization without dedicated sensing hardware

Innovation

• GPU-accelerated AI framework for edge RAN processing
• Plug-and-play dApp architecture on Open RAN platform
• Real-time ISAC using communication signals without hardware changes