arXiv subjects

Zhihui Gao

Publications and source records attributed to Zhihui Gao.

12 recordsLinked to original sources

arXiv2026

Radio-Frequency Convolutional Neural Networks

Running artificial intelligence (AI) models directly on edge devices such as smartphones, wearables, and drones offers low latency, pervasive scalability, and data privacy, but these devices rarely carry the computing capability that modern neural networks demand. Edge accelerators have been developed in response, yet each adds computing hardware to devices already constrained in size, weight, power, and cost (SWaP-C). An alternative lies in what these devices already carry: the frequency mixer in every wireless radio multiplies signals in time, natively performing convolution in the frequency domain. Here we introduce radio-frequency convolutional neural networks (RF-CNNs), which repurpose existing communication hardware for CNN inference. Multi-channel convolutions are mapped onto frequency tones for a passive mixer to execute in a single pass. We experimentally demonstrate that RF-CNN runs deep CNNs up to 26.4 million parameters and nine layers from classification of wireless signals and images to controllable image generation, close to full-precision performance. Because the weights arrive over the air and the analog hardware is shared with communication, the edge device spends energy only on data preparation and readout-down to 0.72 femtojoules per multiply-accumulate, two orders of magnitude less than it would cost on an added digital processor. These results suggest that deployed wireless infrastructure can bring efficient, state-of-the-art AI inference to the billions of devices it already connects.

cs.LG↗

arXiv2026

OASIS: Outlier-Aware LUT-Based GEMM with Dual-Side Quantization for LLM Inference Acceleration

Large language models (LLMs) have demonstrated impressive capabilities across a wide range of applications, but demand substantial memory and compute resources during inference. Existing quantization methods expose a trade-off between efficiency and accuracy: weight-only quantization (WOQ) incurs costly dequantization overheads, while integer weight-and-activation quantization (INT-WAQ) reduces precision and degrades model quality. Non-uniform weight-and-activation quantization (NU-WAQ) can better capture the non-uniform distributions of LLM weights and activations, yet remains incompatible with conventional low-precision compute units. This paper presents OASIS, a lookup table (LUT)-based architecture that enables efficient general matrix multiplication (GEMM) between non-uniformly quantized weights and activations without requiring dequantization. OASIS employs pre-computed Cartesian Product LUTs, achieving a 64x reduction in LUT size and enabling a 1024x higher computational parallelism over existing LUT-based GEMM methods. To preserve accuracy under aggressive activation quantization, OASIS introduces an outlier-aware quantization scheme with concurrent LUT-based GEMM and error compensation for outliers. Furthermore, we design Orizuru, an efficient top-k detection engine for real-time activation outlier identification. According to extensive evaluations, OASIS incurs an average accuracy drop of only 1.98% compared to the FP16 baseline, which is 5.18% lower than Atom. On the hardware side, OASIS achieves an average 3.00x speedup and a 1.44x energy efficiency improvement compared to the FIGLUT accelerator.

cs.LG↗

arXiv2026

FastPFRec: A Fast Personalized Federated Recommendation with Secure Sharing

Graph neural network (GNN)-based federated recommendation systems effectively capture user-item relationships while preserving data privacy. However, existing methods often face slow convergence on graph data and privacy leakage risks during collaboration. To address these challenges, we propose FastPFRec (Fast Personalized Federated Recommendation with Secure Sharing), a novel framework that enhances both training efficiency and data security. FastPFRec accelerates model convergence through an efficient local update strategy and introduces a privacy-aware parameter sharing mechanism to mitigate leakage risks. Experiments on four real-world datasets (Yelp, Kindle, Gowalla-100k, and Gowalla-1m) show that FastPFRec achieves 32.0% fewer training rounds, 34.1% shorter training time, and 8.1% higher accuracy compared with existing baselines. These results demonstrate that FastPFRec provides an efficient and privacy-preserving solution for scalable federated recommendation.

cs.IR↗

arXiv2026

Nexus: Efficient and Scalable Multi-Cell mmWave Baseband Processing with Heterogeneous Compute

The rapid adoption of 5G New Radio (NR), particularly in the millimeter-wave (mmWave) spectrum, imposes stringent demands on the flexibility, scalability, and efficiency of baseband processing. While virtualized Radio Access Networks (vRANs) enable dynamic spectrum sharing across cells, compute resource allocation for baseband processing, especially in multi-cell deployments with heterogeneous workloads, remains underexplored. In this paper, we present NEXUS, the first system to realize real-time, virtualized multi-cell mmWave baseband processing on a single server with heterogeneous compute resources. NEXUS integrates software-based digital signal processing pipelines with hardware-accelerated LDPC decoding, and introduces a novel framework for sharing Intel's ACC100 eASIC across multiple CPU cores via virtual functions (VFs). For single-cell operation, NEXUS employs a random forest (RAF)-based model that predicts the most energy-efficient resource allocation for the given cell configuration with microsecond-level inference latency and high accuracy. For multi-cell scenarios, NEXUS introduces a power-aware scheduler that incorporates a lightweight contention model to adjust resource allocation strategies under concurrent execution. Through extensive evaluation across various Frequency Range 2 (FR2) cell configurations, we show that NEXUS supports up to 16 concurrent cells under full load, achieving 5.37Gbps aggregate throughput, while reducing the multi-cell scheduling search space by orders of magnitude. These results demonstrate that virtualized, resource-aware baseband processing is both practical and efficient for next-generation vRAN systems.

cs.NI↗

arXiv2025

IoT-MCP: Bridging LLMs and IoT Systems Through Model Context Protocol

The integration of Large Language Models (LLMs) with Internet-of-Things (IoT) systems faces significant challenges in hardware heterogeneity and control complexity. The Model Context Protocol (MCP) emerges as a critical enabler, providing standardized communication between LLMs and physical devices. We propose IoT-MCP, a novel framework that implements MCP through edge-deployed servers to bridge LLMs and IoT ecosystems. To support rigorous evaluation, we introduce IoT-MCP Bench, the first benchmark containing 114 Basic Tasks (e.g., ``What is the current temperature?'') and 1,140 Complex Tasks (e.g., ``I feel so hot, do you have any ideas?'') for IoT-enabled LLMs. Experimental validation across 22 sensor types and 6 microcontroller units demonstrates IoT-MCP's 100% task success rate to generate tool calls that fully meet expectations and obtain completely accurate results, 205ms average response time, and 74KB peak memory footprint. This work delivers both an open-source integration framework (https://github.com/Duke-CEI-Center/IoT-MCP-Servers) and a standardized evaluation methodology for LLM-IoT systems.

cs.DC↗

arXiv2025

Chameleon: Integrated Sensing and Communication with Sub-Symbol Beam Switching in mmWave Networks

Next-generation cellular networks are envisioned to integrate sensing capabilities with communication, particularly in the millimeter-wave (mmWave) spectrum, where beamforming using large-scale antenna arrays enables directional signal transmissions for improved spatial multiplexing. In current 5G networks, however, beamforming is typically designed either for communication or sensing (e.g., beam training during link establishment). In this paper, we present Chameleon, a novel framework that augments and rapidly switches beamformers during each demodulation reference signal (DMRS) symbol to achieve integrated sensing and communication (ISAC) in 5G mmWave networks. Each beamformer introduces an additional sensing beam toward target angles while maintaining the communication beams toward multiple users. We implement Chameleon on a 28 GHz software-defined radio testbed supporting over-the-air 5G physical downlink shared channel (PDSCH) transmissions. Extensive experiments in open environments show that Chameleon achieves multi-user communication with a sum data rate of up to 0.80 Gbps across two users. Simultaneously, Chameleon employs a beamformer switching interval of only 0.24 μs, therefore producing a 31x31-point 2D imaging within just 0.875 ms. Leveraging machine learning, Chameleon further enables object localization with median errors of 0.14 m (distance) and 0.24° (angle), and material classification with 99.0% accuracy.

cs.NI↗

arXiv2025

BatStation: Toward In-Situ Radar Sensing on 5G Base Stations with Zero-Shot Template Generation

The coexistence between incumbent radar signals and commercial 5G signals necessitates a versatile and ubiquitous radar sensing for efficient and adaptive spectrum sharing. In this context, leveraging the densely deployed 5G base stations (BS) for radar sensing is particularly promising, offering both wide coverage and immediate feedback to 5G scheduling. However, the targeting radar signals are superimposed with concurrent 5G uplink transmissions received by the BS, and practical deployment also demands a lightweight, portable radar sensing model. This paper presents BatStation, a lightweight, in-situ radar sensing framework seamlessly integrated into 5G BSs. BatStation leverages uplink resource grids to extract radar signals through three key components: (i) radar signal separation to cancel concurrent 5G transmissions and reveal the radar signals, (ii) resource grid reshaping to align time-frequency resolution with radar pulse characteristics, and (iii) zero-shot template correlation based on a portable model trained purely on synthetic data that supports detection, classification, and localization of radar pulses without fine-tuning using experimental data. We implement BatStation on a software-defined radio (SDR) testbed and evaluate its performance with real 5G traffic in the CBRS band. Results show robust performance across diverse radar types, achieving detection probabilities of 97.02% (PUCCH) and 79.23% (PUSCH), classification accuracy up to 97.00%, and median localization errors of 2.68-6.20 MHz (frequency) and 24.6-32.4 microseconds (time). Notably, BatStation achieves this performance with a runtime latency of only 0.11/0.94 ms on GPU/CPU, meeting the real-time requirement of 5G networks.

cs.NI↗

arXiv2025

Machine Intelligence on Wireless Edge Networks

Machine intelligence on edge devices enables low-latency processing and improved privacy, but is often limited by the energy and delay of moving and converting data. Current systems frequently avoid local model storage by sending queries to a server, incurring uplink cost, network latency, and privacy risk. We present the opposite approach: broadcasting model weights to clients that perform inference locally using in-physics computation inside the radio receive chain. A base station transmits weights as radio frequency (RF) waveforms; the client encodes activations onto the waveform and computes the result using existing mixer and filter stages, RF components already present in billions of edge devices such as cellphones, eliminating repeated signal conversions and extra hardware. Analysis shows that thermal noise and nonlinearity create an optimal energy window for accurate analog inner products. Hardware-tailored training through a differentiable RF chain preserves accuracy within this regime. Circuit-informed simulations, consistent with a companion experiment, demonstrate reduced memory and conversion overhead while maintaining high accuracy in realistic wireless edge scenarios.

cs.ET↗

arXiv2025

Disaggregated Deep Learning via In-Physics Computing at Radio Frequency

Modern edge devices, such as cameras, drones, and Internet-of-Things nodes, rely on deep learning to enable a wide range of intelligent applications, including object recognition, environment perception, and autonomous navigation. However, deploying deep learning models directly on the often resource-constrained edge devices demands significant memory footprints and computational power for real-time inference using traditional digital computing architectures. In this paper, we present WISE, a novel computing architecture for wireless edge networks designed to overcome energy constraints in deep learning inference. WISE achieves this goal through two key innovations: disaggregated model access via wireless broadcasting and in-physics computation of general complex-valued matrix-vector multiplications directly at radio frequency. Using a software-defined radio platform with wirelessly broadcast model weights over the air, we demonstrate that WISE achieves 95.7% image classification accuracy with ultra-low operation power of 6.0 fJ/MAC per client, corresponding to a computation efficiency of 165.8 TOPS/W. This approach enables energy-efficient deep learning inference on wirelessly connected edge devices, achieving more than two orders of magnitude improvement in efficiency compared to traditional digital computing.

cs.ET↗

arXiv2024

Geo2SigMap: High-Fidelity RF Signal Mapping Using Geographic Databases

Radio frequency (RF) signal mapping, which is the process of analyzing and predicting the RF signal strength and distribution across specific areas, is crucial for cellular network planning and deployment. Traditional approaches to RF signal mapping rely on statistical models constructed based on measurement data, which offer low complexity but often lack accuracy, or ray tracing tools, which provide enhanced precision for the target area but suffer from increased computational complexity. Recently, machine learning (ML) has emerged as a data-driven method for modeling RF signal propagation, which leverages models trained on synthetic datasets to perform RF signal mapping in "unseen" areas. In this paper, we present Geo2SigMap, an ML-based framework for efficient and high-fidelity RF signal mapping using geographic databases. First, we develop an automated framework that seamlessly integrates three open-source tools: OpenStreetMap (geographic databases), Blender (computer graphics), and Sionna (ray tracing), enabling the efficient generation of large-scale 3D building maps and ray tracing models. Second, we propose a cascaded U-Net model, which is pre-trained on synthetic datasets and employed to generate detailed RF signal maps, leveraging environmental information and sparse measurement data. Finally, we evaluate the performance of Geo2SigMap via a real-world measurement campaign, where three types of user equipment (UE) collect over 45,000 data points related to cellular information from six LTE cells operating in the citizens broadband radio service (CBRS) band. Our results show that Geo2SigMap achieves an average root-mean-square-error (RMSE) of 6.04 dB for predicting the reference signal received power (RSRP) at the UE, representing an average RMSE improvement of 3.59 dB compared to existing methods.

eess.SP↗

arXiv2021

CRISLoc: Reconstructable CSI Fingerprintingfor Indoor Smartphone Localization

Channel state information (CSI) based fingerprinting for WIFI indoor localization has attracted lots of attention very recently.The frequency diverse and temporally stable CSI better represents the location dependent channel characteristics than the coarsereceived signal strength (RSS). However, the acquisition of CSI requires the cooperation of access points (APs) and involves only dataframes, which imposes restrictions on real-world deployment. In this paper, we present CRISLoc, the first CSI fingerprinting basedlocalization prototype system using ubiquitous smartphones. CRISLoc operates in a completely passive mode, overhearing thepackets on-the-fly for his own CSI acquisition. The smartphone CSI is sanitized via calibrating the distortion enforced by WiFi amplifiercircuits. CRISLoc tackles the challenge of altered APs with a joint clustering and outlier detection method to find them. A novel transferlearning approach is proposed to reconstruct the high-dimensional CSI fingerprint database on the basis of the outdated fingerprintsand a few fresh measurements, and an enhanced KNN approach is proposed to pinpoint the location of a smartphone. Our studyreveals important properties about the stability and sensitivity of smartphone CSI that has not been reported previously. Experimentalresults show that CRISLoc can achieve a mean error of around 0.29m in a6m times 8mresearch laboratory. The mean error increases by 5.4 cm and 8.6 cm upon the movement of one and two APs, which validates the robustness of CRISLoc against environment changes.

cs.NI↗

arXiv2021

Hermes: Decentralized Dynamic Spectrum Access System for Massive Devices Deployment in 5G

With the incoming 5G network, the ubiquitous Internet of Things (IoT) devices can benefit our daily life, such as smart cameras, drones, etc. With the introduction of the millimeter-wave band and the thriving number of IoT devices, it is critical to design new dynamic spectrum access (DSA) system to coordinate the spectrum allocation across massive devices in 5G. In this paper, we present Hermes, the first decentralized DSA system for massive devices deployment. Specifically, we propose an efficient multi-agent reinforcement learning algorithm and introduce a novel shuffle mechanism, addressing the drawbacks of collision and fairness in existing decentralized systems. We implement Hermes in 5G network via simulations. Extensive evaluations show that Hermes significantly reduces collisions and improves fairness compared to the state-of-the-art decentralized methods. Furthermore, Hermes is able to adapt the environmental changes within 0.5 seconds, showing its deployment practicability in dynamic environment of 5G.

cs.NI↗