About

Tingjun Chen is an assistant professor in the Department of Electrical and Computer Engineering at Duke University, affiliated with the Pierre Lamond Department of Electrical and Computer Engineering. His research interests include wireless networking, next-generation wireless and optical networks, edge cloud and computing, Internet-of-Things (IoT), mobile and embedded systems, platforms, and testbeds. He has a strong academic background with a B.S. from Tsinghua University in China, an M.S. and Ph.D. from Columbia University, and postdoctoral training at Yale University. Chen has received numerous awards and honors, including multiple NSF CAREER Awards, the Corning Outstanding Student Paper Competition Finalist, and the IBM Academic Award. His teaching includes courses on IoT systems, advanced topics in electrical and computer engineering, and projects related to wireless and computer science. His contributions span various high-impact publications and innovative research in wireless communication, IoT, and network systems.

Research topics

• Computer Science
• Political Science
• Computer network
• Telecommunications
• Psychology
• Distributed computing
• Neuroscience
• Electronic engineering
• Engineering
• Operating system
• Mathematics
• Biology
• Electrical engineering
• Public relations
• Medicine
• Marketing
• Social psychology
• Computer architecture
• Immunology
• Pathology
• Business
• Mathematical optimization
• Embedded system

Selected publications

• RISE: Real-time Image Processing for Spectral Energy Detection and Localization

  ArXiv.org · 2026-03-20

  articleOpen accessSenior author

  Energy detection is widely used for spectrum sensing, but accurately localizing the time and frequency occupation of signals in real-time for efficient spectrum sharing remains challenging. To address this challenge, we present RISE, a software-based spectrum sensing system designed for real-time signal detection and localization. RISE treats time-frequency spectrum plots as images and applies adaptive thresholding, morphological operations, and connected component labeling with a multi-threaded architecture. We evaluate RISE using both synthetic data and controlled over-the-air (OTA) experiments across diverse signal types. Results show that RISE satisfies real-time latency constraints while achieving a probability of detection of 80.42% at an intersection-over-union (IoU) threshold of 0.4. RISE sustains a raw I/Q input rate of 3.2 Gbps for 100 MHz bandwidth sensing with time and frequency resolutions of 10.24 us and 97.6 kHz, respectively. Compared to Searchlight, a representative energy-based method, RISE achieves 20.51x lower latency and 22.31% higher IoU. Compared to machine learning baselines, RISE improves IoU by 56.02% over DeepRadar while meeting the real-time deadline, which a GPU-accelerated U-Net exceeds by 213.38x.

  PublisherOA PDF

• Disaggregated machine learning via in-physics computing at radio frequency

  Science Advances · 2026-01-09 · 1 citations

  articleOpen accessSenior authorCorresponding

  Modern edge devices, such as cameras, drones, and internet-of-things nodes, rely on machine learning to enable a wide range of intelligent applications. However, deploying machine learning models directly on the often resource-constrained edge devices demands substantial memory footprints and computational power for real-time inference using traditional digital computing architectures. In this paper, we present WISE, computing architecture for wireless edge networks with two key innovations: disaggregated model access via over-the-air wireless broadcasting for simultaneous inference on multiple edge devices, and in-physics computation of general complex-valued matrix-vector multiplications directly at radio frequency driven by a single frequency mixer. Using a software-defined radio platform, WISE achieves 95.7% image classification accuracy (97.2% audio classification accuracy) with ultralow energy consumption of 6.0 fJ/MAC (2.8 fJ/MAC), which is more than 10× improvement compared to traditional digital computing, e.g., on modern GPUs.

  PublisherDOI

• Refining responsive antimicrobial nanomaterials from random peptide libraries via machine learning

  Responsive materials · 2026-02-01 · 2 citations

  articleOpen access

  Abstract Antimicrobial resistance has evolved into one of the most serious threats to global public health, yet generalizable routes for refining random peptide mixtures (RPMs) into defined, stimuli‐responsive, and low‐cost antimicrobial formulations remain limited. Here, a refinement framework is presented. It centers on machine learning and co‐assembly that converts broad‐spectrum RPMs into interpretable antimicrobial peptide cocktails without exhaustive screening. Specifically, starting from a 10‐mer Arg/Leu RPM (RL0.5), its antimicrobial activity and self‐assembly are quantified, and machine learning is used to prioritize key functional peptides. Leveraging synergistic and co‐assembly behaviors, an optimal combination (AEP) is selected. The resulting defined formulation, RL10, achieves a fourfold increase in in vitro activity against Escherichia coli and exhibits a reduced critical aggregation concentration relative to the starting RPM. Overall, this study presents a practical path from complex, low‐cost precursors to efficient, co‐assembling antimicrobial cocktails, and summarizes explainable design rules that support engineering and industrialization.

  PublisherOA PDFDOI

• Phantora: Maximizing Code Reuse in Simulation-based Machine Learning System Performance Estimation

  ArXiv.org · 2025-05-02

  preprintOpen access

  Modern machine learning (ML) training workloads place substantial demands on both computational and communication resources. Consequently, accurate performance estimation has become increasingly critical for guiding system design decisions, such as the selection of parallelization strategies, cluster configurations, and hardware provisioning. Existing simulation-based performance estimation requires reimplementing the ML framework in a simulator, which demands significant manual effort and is hard to maintain as ML frameworks evolve rapidly. This paper introduces Phantora, a hybrid GPU cluster simulator designed for performance estimation of ML training workloads. Phantora executes unmodified ML frameworks as is within a distributed, containerized environment. Each container emulates the behavior of a GPU server in a large-scale cluster, while Phantora intercepts and simulates GPU- and communication-related operations to provide high-fidelity performance estimation. We call this approach hybrid simulation of ML systems, in contrast to traditional methods that simulate static workloads. The primary advantage of hybrid simulation is that it allows direct reuse of ML framework source code in simulation, avoiding the need for reimplementation. Our evaluation shows that Phantora provides accuracy comparable to static workload simulation while supporting three state-of-the-art LLM training frameworks out-of-the-box. In addition, Phantora operates on a single GPU, eliminating the need for the resource-intensive trace collection and workload extraction steps required by traditional trace-based simulators. Phantora is open-sourced at https://github.com/QDelta/Phantora.

  PublisherOA PDFDOI

• Generalized few-shot transfer learning architecture for modeling the EDFA gain spectrum

  Journal of Optical Communications and Networking · 2025-06-05 · 2 citations

  articleOpen accessCorresponding

  Accurate modeling of the gain spectrum in erbium-doped fiber amplifiers (EDFAs) is essential for optimizing optical network performance, particularly as networks evolve toward multi-vendor solutions. In this work, we propose a generalized few-shot transfer learning architecture based on a semi-supervised self-normalizing neural network (SS-NN) that leverages internal EDFA features—such as VOA input/output power and attenuation—to improve gain spectrum prediction. Our SS-NN model employs a two-phase training strategy comprising unsupervised pre-training with noise-augmented measurements and supervised fine-tuning with a custom-weighted MSE loss. Furthermore, we extend the framework with transfer learning (TL) techniques that enable both homogeneous (same-feature space) and heterogeneous (different-feature sets) model adaptation across booster, pre-amplifier, and ILA EDFAs. To address feature mismatches in heterogeneous TL, we incorporate a covariance matching loss to align second-order feature statistics between the source and target domains. Extensive experiments conducted across 26 EDFAs in the COSMOS and Open Ireland testbeds demonstrate that the proposed approach significantly reduces the number of measurement requirements on the system while achieving lower mean absolute errors and improved error distributions compared to benchmark methods.

  PublisherOA PDFDOI

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

  2025-10-28 · 2 citations

  articleOpen access

  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://llm4iot.org) and a standardized evaluation methodology for LLM-IoT systems.

  PublisherDOI

• Design of an Intelligent Diagnosis and Self-Healing System for ZigBee Networks Based on Machine Learning

  2025-05-23

  articleSenior author

  Aiming at the issues of signal attenuation, electromagnetic interference, and low maintenance efficiency in ZigBee networks in complex indoor scenarios, this paper proposes an intelligent diagnosis and self-healing system based on machine learning. By innovatively integrating network-layer feature aggregation technology, lightweight machine learning algorithms, and topology visualization diagnosis systems, the system constructs a spatiotemporal analysis model using the inherent data of the ZigBee protocol stack. It employs a dual-model architecture of Random Forest and 1D-CNN to classify faults and provides a visual diagnosis interface through dynamic topology mapping technology. Experiments show that the system can achieve a fault location accuracy rate of 100% and reduce maintenance costs by 42%. It supports device migration across gateways (with a success rate of ≥99.5%) and dynamic optimization of network topology, with a network recovery time of <3 minutes. This system offers a non-contact maintenance solution for smart home and industrial Internet of Things scenarios, significantly enhancing the robustness and maintainability of large-scale ZigBee networks.

  PublisherDOI

• A Generalized Deep Learning Model for Signal Coverage Prediction in the CBRS Band

  2025-05-12 · 2 citations

  articleSenior author

  In cellular networks, received signal strength (RSS) prediction plays an essential role in cellular network planning and deployment, as it aims to estimate the wireless signal quality that a base station (BS) can deliver to user equipment (UE) within an area of interest. While there have been extensive works on developing analytical and empirical channel models for signal coverage prediction, these models typically do not consider cell-specific environmental information such as building footprints and other types of clutters. As a result, their performance in RSS prediction can deviate from real-world deployments and measurements. In this paper, we bridge such a gap by implementing state-of-the-art RSS prediction methods based on deep learning (DL) and with evaluations using real-world RSS measurements collected from an LTE network operating in the Citizens Broadband Radio Service (CBRS) band. We also present a comprehensive comparison of the RSS prediction performance compared to analytical and empirical channel models as well as ray tracing (RT) methods. Our evaluations reveal that the existing empirical/analytical channel models and RT methods exhibit unstable RSS prediction performance depending on the environments, with maximum root mean square error (RMSE) values ranging from 7.12–12.43dB and 6.61–18.96dB, respectively. In contrast, the DL method outperforms these baseline methods, achieving a more stable performance with RMSE values ranging between 5.71–12.18 dB across the 10 PCIs, therefore demonstrating its robustness and generalizability.

  PublisherDOI

• RaGNNarok: A Light-Weight Graph Neural Network for Enhancing Radar Point Clouds on Unmanned Ground Vehicles

  ArXiv.org · 2025-07-01

  preprintOpen access

  Low-cost indoor mobile robots have gained popularity with the increasing adoption of automation in homes and commercial spaces. However, existing lidar and camera-based solutions have limitations such as poor performance in visually obscured environments, high computational overhead for data processing, and high costs for lidars. In contrast, mmWave radar sensors offer a cost-effective and lightweight alternative, providing accurate ranging regardless of visibility. However, existing radar-based localization suffers from sparse point cloud generation, noise, and false detections. Thus, in this work, we introduce RaGNNarok, a real-time, lightweight, and generalizable graph neural network (GNN)-based framework to enhance radar point clouds, even in complex and dynamic environments. With an inference time of just 7.3 ms on the low-cost Raspberry Pi 5, RaGNNarok runs efficiently even on such resource-constrained devices, requiring no additional computational resources. We evaluate its performance across key tasks, including localization, SLAM, and autonomous navigation, in three different environments. Our results demonstrate strong reliability and generalizability, making RaGNNarok a robust solution for low-cost indoor mobile robots.

  PublisherOA PDFDOI

• On the Optimization and Stability of Sectorized Wireless Networks

  IEEE Transactions on Networking · 2025-07-04

  article

  Future wireless networks need to support the increasing demands for high data rates and improved coverage. One promising solution is sectorization, where an infrastructure node is equipped with multiple sectors employing directional communication. Although the concept of sectorization is not new, it is critical to fully understand the potential of sectorized networks, such as the rate gain achieved when multiple sectors can be simultaneously activated. In this paper, we focus on sectorized wireless networks, where sectorized infrastructure nodes with beam-steering capabilities form a multi-hop mesh network. We present a sectorized node model and characterize the capacity region of these sectorized networks. We define the flow extension ratio and the corresponding sectorization gain, which quantitatively measure the performance gain introduced by node sectorization as a function of the network flow. Our objective is to find the sectorization of each node that achieves the maximum flow extension ratio, and thus the sectorization gain. Towards this goal, we formulate the corresponding optimization problem and develop an efficient distributed algorithm that obtains the node sectorization under a given network flow with an approximation ratio of 2/3. Additionally, we emphasize the class of Even Homogeneous Sectorizations, which simultaneously enhances the efficiency of dynamic routing schemes with unknown arrival rates and increases network capacity. We further propose that if sectorization can be adapted dynamically over time, either a backpressure-driven or maximum weighted b-matching-based routing approach can be employed, thereby expanding the achievable capacity region while preserving stability under unknown traffic conditions. Through extensive simulations, we evaluate the sectorization gain and the performance of the proposed algorithms in various network scenarios.

  PublisherDOI

Recent grants

• Collaborative Research: SWIFT: SHIELD: A Software-Hardware Approach for Spectrum Coexistence with Rapid Interferer Learning, Detection, and Mitigation

  NSF · $258k · 2021–2025

• Collaborative Research: CNS Core: Medium: Softwarizing Millimeter-wave Radio Access Networks (RANs) at the Edge

  NSF · $550k · 2022–2026

• EAGER: An Integrated Fiber Sensing and Communication Living Lab in the Research Triangle

  NSF · $300k · 2023–2026

• Collaborative Research: SII-NRDZ:Spectrum Sharing via Consumption Models and Telemetry - Prototyping and Field Testing in an Urban FCC Innovation Zone

  NSF · $190k · 2023–2026

Frequent coauthors

• Gil Zussman

  Columbia University

  187 shared

• Harish Krishnaswamy

  Columbia University

  156 shared

• Mahmood Baraani Dastjerdi

  139 shared

• Aravind Nagulu

  Washington University in St. Louis

  134 shared

• Negar Reiskarimian

  Massachusetts Institute of Technology

  130 shared

• Aditya Gaonkar

  Columbia University

  127 shared

• Sohail Ahasan

  Columbia University

  124 shared

• Armagan Dascurcu

  Columbia University

  121 shared

Labs

• Pierre R. Lamond Department of Electrical and Computer EngineeringPI

Education

• Ph.D., Electrical Engineering and Computer Sciences

  University of California, Berkeley

  2012

• M.S., Electrical Engineering and Computer Sciences

  University of California, Berkeley

  2008

• B.S., Electronic Information Engineering

  University of Science and Technology of China

  2006

Awards & honors

• National Science Foundation CAREER Awards - Multiple (2025)
• Corning Outstanding Student Paper Competition Finalist (2024…
• Top-Scored Paper IEEE/Optica OFC'24 (2024)
• IBM Academic Award (2023)
• Highest Scoring Paper ECOC'23 (2023)