About

Kevin Fu is a professor at Northeastern University, affiliated with the College of Engineering and the Khoury College of Computer Sciences. His research focuses on protecting emerging sensor technology in biomedical engineering and cyber-physical systems, with particular emphasis on cybersecurity in healthcare and medical devices. He is involved in establishing the Archimedes Center for Healthcare and Device Security, which aims to help manufacturers and industry experts navigate cybersecurity hazards and meet FDA requirements. Fu has led significant initiatives such as the UPGRADE project, a national effort to create hospital-scale digital twins for cybersecurity vulnerability discovery and remediation, supported by ARPA-H. He also directs a multidisciplinary team awarded a NSF grant to secure neural implants and brain-controlled interfaces, emphasizing resilience and safety engineering. Fu has served as the first director of medical device security at the Food and Drug Administration and has addressed cybersecurity risks in medical devices at congressional hearings. His contributions include advancing cybersecurity measures for medical devices, addressing vulnerabilities in healthcare infrastructure, and developing techniques to detect and prevent cyber threats, making him a leading figure in cybersecurity for healthcare technology.

Research topics

• Computer Science
• Computer Security
• Artificial Intelligence
• Computer hardware
• Business
• Computer vision
• Risk analysis (engineering)
• Electrical engineering
• Software engineering
• World Wide Web
• Internet privacy
• Engineering
• Telecommunications
• Operating system

Selected publications

• ARMOUR US: Android Runtime Zero-permission Sensor Usage Monitoring from User Space

  2025-06-27

  articleOpen accessSenior author

  This work investigates how to monitor access to Android zero-permission sensors which could cause privacy leakage to users. Moreover, monitoring such sensitive access allows security researchers to characterize potential sensor abuse patterns. Zero-permission sensors such as accelerometers have become an indispensable part of Android devices. The critical information they provide has attracted extensive research investigating how data collectors could capture more sensor data to enable both benign and exploitative applications. In contrast, little work has explored how to enable data providers, such as end users, to understand sensor usage. While existing methods such as static analysis and hooking-based dynamic analysis face challenges of requiring complicated development chains, rooting privilege, and app-specific reverse engineering analysis, our work aims to bridge this gap by developing ARMOUR for user-space runtime monitoring, leveraging the intrinsic sampling rate variation and convergence behaviors of Android. ARMOUR enables privacy-aware users to easily monitor how third-party apps use sensor data and support security researchers to perform rapid app-agnostic sensor access analysis. Our evaluation with 1,448 commercial applications shows the effectiveness of ARMOUR in detecting sensor usage in obfuscated code and other conditions, and observes salient sensor abuse patterns such as 50% of apps from seemingly sensor-independent categories accessing data of multiple zero-permission sensors. We analyze the impact of Android's recent policy changes on zero-permission sensors and remaining technical and regulatory problems.

  PublisherDOI

• Eyehearyou: Probing Location Identification Via Occluded Smartphone Cameras and Ultrasound

  2025-05-05

  articleSenior author

  This paper explores how to localize a device equipped solely with camera sensors by leveraging the unintended response of occluded camera hardware to transmitted ultrasound-specifically, determining with high probability which ultrasound transmission pattern was injected during image capture, based on distinctive ultrasound signals unwittingly detected by the image sensor. Prior device location identification methods require the use of dedicated hardware or protocols, e.g., microphone arrays or GPS. We envision a new potential mechanism by leveraging ubiquitous camera hardware on mobile and IoT devices to receive acoustic signals produced by ultrasonic localization beacons, even when the camera is occluded. We discover that ultrasonic signals affect gyroscopes integral to modern camera sensors' stabilization hardware and induce distinct destabilization signals in the dark images captured by occluded cameras. This work provides theoretical analysis, simulation modeling, and experimental evidence of how this optical acoustic side channel creates different noise patterns in camera images when the camera is subjected to different ultrasound stimuli. Our evaluation with 119 videos captured by a smartphone camera over multiple days shows success in detecting whether the smartphone is near an ultrasonic transmitter that can be associated with different locations.

  PublisherDOI

• How Lasers Exploit Photoacoustic and Photoelectric Phenomena to Inject Signals into MEMS Microphones

  Journal of Hardware and Systems Security · 2025-05-01

  articleOpen accessSenior author

  Abstract An amplitude-modulated laser can be used to generate false, yet coherent acoustic signals on the outputs of MEMS microphones. While this vulnerability has ramifications on the security of cyber-physical systems that trust these microphones, the physical explanation of this effect remained a mystery. Without an understanding of the physical phenomena contributing to this signal injection, it is difficult to design effective and reliable defenses. In this work, we show the degree to which the mechanisms of thermoelastic bending, thermal diffusion, and photocurrent generation are used to inject signals into MEMS microphones. We provide models for each of these mechanisms, develop a procedure to empirically determine their relative contributions, and highlight the effects on eight commercial MEMS microphones. We accomplish this with a precise setup to isolate each mechanism using several laser wavelengths and a vacuum chamber. The results indicate that the injected signal on the microphone is dependent on the wavelength of the incoming light. Shorter wavelengths (such as a 450 nm blue laser) exploit photoacoustic effects, and the periodic heating and expansion of air is the dominant factor in seven of eight sample microphones. Longer wavelengths (such as a 904 nm infrared laser) exploit photoelectric effects on the sensitive ASIC, generating signals that are between 2x and 100x stronger than photoacoustic signals in six of eight sample microphones. This understanding of the physical causality of laser signal injection leads to recommendations for future laser-resistant microphone designs. These include adding light-blocking structures at the system or device level, improving to glob top application, and adding simple light or temperature sensors for injection detection. Based on the fundamental causality, we also suggest potential vulnerabilities within other sensors with similar characteristics to MEMS microphones, such as conventional microphones, ultrasonic sensors, and inertial sensors.

  PublisherOA PDFDOI

• Threats to Patient Safety From Cybersecurity Flaws—A New Never Event

  JAMA · 2025-07-07 · 1 citations

  articleSenior author

  This Viewpoint discusses medical device cybersecurity vulnerabilities and the threat they pose to patient safety.

  PublisherDOI

• Promoting the Resilience of Health Care Information Systems—The Day Hospitals Stood Still

  JAMA Health Forum · 2024-11-27 · 2 citations

  articleOpen accessSenior author

  This Viewpoint describes health care disruption owing to a global computer system outage and advocates for preventive methods to avoid future IT events disrupting the health care sector.

  PublisherOA PDFDOI

• From Virtual Touch to Tesla Command: Unlocking Unauthenticated Control Chains From Smart Glasses for Vehicle Takeover

  2024-05-19 · 7 citations

  article

  This paper studies vulnerabilities at the intersection of wearable devices and automated control systems. Particularly, we focus on exploiting smart glasses as an entry point and unveil the threats of taking over security-critical automated control chains without user verification or interaction. These vulnerabilities can be especially pertinent in scenarios where security mechanisms only depend on entry point security with minimal user verification (relying on complete trust over previous nodes in automated control chains). We have validated the effects of our attacks on real-world systems (e.g., Tesla vehicles) that are controlled by software and automation tools such as Apple Shortcuts or IFTTT. We show how our contactless, speaker-independent, and electromagnetic interference based attacks can control functionalities such as unlocking doors and initiating remote start of Tesla vehicles, even though the victim’s phone is in a lock-screen status. Our findings not only demonstrate the potential for unauthorized control over automated, connected systems but also highlight the urgent need for more robust security measures in the integration of wearable technology with broader automation frameworks.

  PublisherDOI

• How Lasers Exploit Photoacoustic and PhotoelectricPhenomena to Inject Signals into MEMS Microphones

  Research Square · 2024-04-11 · 2 citations

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• Hybrid Threat Detection Architecture for Automotive Networks: Integrating AI-Based IDS with Lightweight Encryption for Secure V2X and In-Vehicle Communication

  2024-01-01

  articleOpen access1st authorCorresponding
  PublisherDOI

• EM Eye: Characterizing Electromagnetic Side-channel Eavesdropping on Embedded Cameras

  2024-01-01 · 19 citations

  articleOpen accessSenior author

  IoT devices and other embedded systems are increasingly equipped with cameras that can sense critical information in private spaces.The data security of these cameras, however, has hardly been scrutinized from the hardware design perspective.Our paper presents the first attempt to analyze the attack surface of physical-channel eavesdropping on embedded cameras.We characterize EM Eye-a vulnerability in the digital image data transmission interface that allows adversaries to reconstruct high-quality image streams from the cameras' unintentional electromagnetic emissions, even from over 2 meters away in many cases.Our evaluations of 4 popular IoT camera development platforms and 12 commercial off-the-shelf devices with cameras show that EM Eye poses threats to a wide range of devices, from smartphones to dash cams and home security cameras.By exploiting this vulnerability, adversaries may be able to visually spy on private activities in an enclosed room from the other side of a wall.We provide root cause analysis and modeling that enable system defenders to identify and simulate mitigation against this vulnerability, such as improving embedded cameras' data transmission protocols with minimum costs.We further discuss EM Eye's relationship with known computer display eavesdropping attacks to reveal the gaps that need to be addressed to protect the data confidentiality of sensing systems.

  PublisherDOI

• GhostType: The Limits of Using Contactless Electromagnetic Interference to Inject Phantom Keys into Analog Circuits of Keyboards

  2024-01-01 · 6 citations

  articleOpen access

  Keyboards are the primary peripheral input devices for various critical computer application scenarios.This paper performs a security analysis of the keyboard sensing mechanisms and uncovers a new class of vulnerabilities that can be exploited to induce phantom keys-fake keystrokes injected into keyboards' analog circuits in a contactless way using electromagnetic interference (EMI).Besides regular keystrokes, such phantom keys also include keystrokes that human operators cannot achieve, such as rapidly injecting over 10,000 keys per minute and injecting hidden keys that do not exist on the physical keyboard.The underlying principles of phantom key injections consist in inducing false voltages on keyboard sensing GPIO pins through EMI coupled onto matrix circuits.We investigate the voltage and timing requirements of injection signals both theoretically and empirically to establish the theory of phantom key injection.To validate the threat of keyboard sensing vulnerabilities, we design GhostType that can cause denial-of-service of the keyboard and inject random keystrokes as well as certain targeted keystrokes of the adversary's choice.We have validated GhostType on 48 of 50 off-the-shelf keyboards/keypads from 20 brands, including both membrane/mechanical structures and USB/Bluetooth protocols.Some example consequences of GhostType include completely blocking keyboard operations, crashing and turning off downstream computers, and deleting computer files.Finally, we glean lessons from our investigations and propose countermeasures, including shielding keyboards with metal materials and enhancing the keystroke sensing mechanism.

  PublisherDOI

Recent grants

• CT-ISG Collaborative Research: New directions and applications of proxy re-cryptography

  NSF · $79k · 2007–2010

• TWC: Frontier: Collaborative: Enabling Trustworthy Cybersystems for Health and Wellness

  NSF · $2.2M · 2013–2021

• CT-T Collaborative Research: Security for Smart Tags

  NSF · $826k · 2006–2011

• RAPID: SaTC: COVID19: Science of using wirelessly powered sensors to quickly scale up verifiable decontamination of individual N95 respirator masks

  NSF · $200k · 2020–2021

• CAREER: Computational RFID for Securing Zero-Power Pervasive Devices

  NSF · $464k · 2013–2015

Frequent coauthors

• Benjamin Ransford

  University of Washington

  21 shared

• Long Yan

  Yulin University

  21 shared

• Sara Rampazzi

  20 shared

• Wenyuan Xu

  Zhejiang University

  12 shared

• Shane S. Clark

  RTX (United States)

  11 shared

• Mastooreh Salajegheh

  University of Virginia

  11 shared

• Wayne Burleson

  University of Massachusetts Amherst

  10 shared

• Jacob Sorber

  Clemson University

  10 shared

Awards & honors

• 2023 Association for the Advancement of Medical Instrumentat…
• 2022 American Association of Advancement of Science (AAAS) F…
• 2022 Association of Computing Machinery (ACM) Fellow
• 2018 IEEE Fellow
• 2013 IEEE Senior Member