Donghui Dai

Dr. Donghui Dai

Department of Computing, The Hong Kong Polytechnic University, Hong Kong, China
orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 2023
Verified
Conference Paper
Closed access
Donghui DaiDonghui Dai
Donghui Dai
ZAZhenlin An
Zhenlin An
LYLei Yang
Lei Yang
Recent works demonstrated that speech recognition systems or voice assistants can be manipulated by malicious voice commands, which are injected through various inaudible media, such as ultrasound, laser, and electromagnetic interference (EMI). In this work, we explore a new kind of inaudible voice attack through the magnetic interference induced by a wireless charger. Essentially, we show that the microphone components of smart devices suffer from severe magnetic interference when they are enjoying wireless charging, due to the absence of effective protection against the EMI at low frequencies (100 kHz or below). By taking advantage of this vulnerability, we design two inaudible voice attacks, HeartwormAttack and ParasiteAttack, both of which aim to inject malicious voice commands into smart devices being wirelessly charged. They make use of a compromised wireless charger or accessory equipment (called parasite) to inject the voice, respectively. We conduct extensive experiments with 17 victim devices (iPhone, Huawei, Samsung, etc.) and 6 types of voice assistants (Siri, Google STT, Bixby, etc.). Evaluation results demonstrate the feasibility of two proposed attacks with commercial charging settings.

10 July 2023
Verified
Conference Paper
Open access
Donghui DaiDonghui Dai
Donghui Dai
ZAZhenlin An
Zhenlin An
QPQingrui Pan
Qingrui Pan
LYLei Yang
Lei Yang

29 May 2024
Verified
Conference Paper
Open access
ZGZheng Gong
Zheng Gong
ZAZhenlin An
Zhenlin An
Donghui DaiDonghui Dai
Donghui Dai
JTJingyu Tong
Jingyu Tong
SLShuijie Long
Shuijie Long
LYLei Yang
Lei Yang
Within the burgeoning 6G wireless network landscape, there is an intensified push toward achieving all-encompassing accessibility through integrated solutions spanning a multitude of domains. Notwithstanding recent advancements, the conventional relay-centric communication paradigms grapple with scalability and optimal performance issues. In this paper, we introduce MeAnt ---a versatile IoT platform uniquely architected to foster seamless cross-medium communication by leveraging the compact design of piezoelectric-based mechanical antennas (Piezo-MAs). By capitalizing on the propagation attributes of medium-frequency radios emitted from Piezo-MAs, MeAnt promises communication across diverse environments such as air, water, soil, concrete, and even biological tissue, all while maintaining a compact antenna footprint. Moreover, in light of challenges such as potential interference from AM broadcasts and the intrinsic unidirectional nature of Piezo-MAs, we have developed a finely crafted full-stack communication protocol. Comprehensive tests underscore the system's proficiency, demonstrating a penetration depth of up to 10 m in cross-medium environments and realizing a throughput of 8.7 kbps.

4 December 2024
Verified
Conference Paper
Open access
SWShen Wang
Shen Wang
XZXiaopeng Zhao
Xiaopeng Zhao
Donghui DaiDonghui Dai
Donghui Dai
LYLei Yang
Lei Yang
In the era of ubiquitous short-video sharing, an overlooked yet significant privacy risk has arisen: the accidental disclosure of confidential information through reflective surfaces, such as mirrors, glass, or even polished metal. Such reflections can inadvertently disclose sensitive personal details to a broad audience without the awareness of content creators. Our examination of 100 top-rated TikTok short videos reveals that, on average, 37.2% of frames in each video feature identifiable reflective surfaces, posing potential privacy risks. In this work, we introduce a framework designed to scrutinize reflective privacy risks in glass-laden short videos. At the heart of the framework is the development of a reflection-specific neural radiance field, termed RP-NeRF, which enables reflection-aware ray tracing for precise extraction and reconstruction of the reflective scenes from the surfaces they appear on. A detailed field study on the framework indicates that the precision in detecting human presence and recognizing objects from the reconstructed reflective images reaches as high as 90.8% and 89.6%, respectively, even when dealing with a reflective surface that boasts 90% transparency and a mere 4% reflectance rate. These findings highlight the urgent need for greater awareness and advanced solutions to safeguard privacy in our digital age, especially in light of the significant impact of short-video sharing.

4 December 2025 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
YFYuanhao Feng
Yuanhao Feng
Donghui DaiDonghui Dai
Donghui Dai
Jinyang HuangJinyang Huang
Jinyang Huang
PYPanlong Yang
Panlong Yang
XLXiang-Yang Li
Xiang-Yang Li
LYLei Yang
Lei Yang
Accurately monitoring micron-level vibrations with sub-hertz frequency estimation error is critical for early fault detection in industrial equipment. Existing solutions either rely on powered sensors or suffer from limited accuracy in passive operation, restricting scalability and long-term deployment. We present Vibro-Stethos, a fully battery-free sensing system that accurately captures micron-level vibrations with sub-hertz frequency estimation error. It employs a dual-junction field-effect transistor (JFET) analog frontend to convert vibration into impedance modulation and encodes this onto passive RFID backscatter. An embedded RFID chip enables selective tag activation and provides path-invariant reference amplitude normalization. A Graph Attention Network (GAT)-based model adaptively fuses features from spatially distributed tags, enabling robust fault classification under tag sparsity and placement variation. Extensive evaluation demonstrates that Vibro-Stethos achieves amplitude measurement errors within 2 ${\mu }$m, frequency estimation errors below 0.1 Hz, and vibration fault classification accuracy of 93.7%. Real-world deployments on transformers further confirm its diagnostic capability. Vibro-Stethos offers a practical, robust, and accurate battery-free solution for pervasive industrial vibration monitoring.

28 June 2024 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
Donghui DaiDonghui Dai
Donghui Dai
ZAZhenlin An
Zhenlin An
QPQingrui Pan
Qingrui Pan
LYLei Yang
Lei Yang
Mobile payments have grown significantly recently, driven by their contactless feature that minimizes COVID-19 transmission risks. While NFC offers more security and convenience than barcodes and benefits those with amblyopia, many smartphones lack NFC due to module shortages or security decisions. In this work, we present ${\sf MagCode}$ , an innovative method connecting NFC readers with cameras, allowing users to enjoy NFC payment security using prevalent camera technology. At the heart of ${\sf MagCode}$ is the harmless magnetic interference on the CMOS image sensor of a smartphone placed nearby the NFC reader, resulting in a group of barcode-like stripes appearing on the captured images. We take advantage of these stripes to encode the data and achieve simplex communication from an NFC reader to an NFC-denied or NFC-disabled smartphone. In particular, we developed a comprehensive suite of protocols spanning from the physical layer to the transport layer, and we rigorously tested our proof-of-concept prototype on 11 different smart devices. Our extensive evaluations showcase a maximum throughput of 2.58 kbps–surpassing magnetometer-based alternatives by a factor of 58–and demonstrate an average data exchange time of 1.3 seconds for mobile payment transactions between an NFC reader and a smartphone.