Stress Quantification and Monitoring with Wearables

ACM MobiCom'24 Workshop

Catherine Lloyd*, Loic Lorente Lemoine*, Reiyan Al-Shaikh, Kim Tien Ly, Hakan Kayan, Charith Perera, Nhat Pham.

We explore how stress can be quantified and monitored with machine learning and wearable hardware. Our first work [1] looks at fine-tuning a foundation model on an open stress dataset where 20 participants undergo stress-inducing tasks.

Energising wearable and implantable devices through the body channel for healthcare applications

Royal Society Scientific Meeting on the Cyborg Future, June 2025

Kha Huynh, Loic Lemoine, Issac Kwong, Bo Hou, Katarzyna Stawarz, Nhat Pham

We explore a method to achieve reliable body-coupled power transmission for energising healthcare devices placed at arbitrary locations on the body, using the body itself as a low-loss conductive channel. By adapting to variations in transmission conditions during everyday movements, the method enables flexible placement of both transmitter and receiver. Specifically, we (1) characterise the variability in body-coupled power transmission across different postures and during contact with various surrounding objects, (2) design an adaptive mechanism to stabilise delivery, and (3) implement and evaluate a prototype that demonstrates robust, position-independent power transmission.

Edge deployed machine learning for vehicular attacks with online learning and continual learning techniques

MobiUK 2025

Loic Lemoine, Amanjot Kaur, Nhat Pham, Omer Rana

This work presents how anomaly detection for vehicular data, such as EV power consumption and CAN bus data, can undergo online learning on the edge to adapt to new attacks and/or improve model performance. For this study, we present two pieces of work: (1) Pi-based continual learning for EV charger infrastructure attacks, and (2) an Arduino-based CAN bus IDS model with online learning of benign samples.

Seizure Monitoring with Wearables

MobiUK 2024

Loic Lemoine, Nhat Pham

We explore wearable hardware alongside efficient computing for the monitoring of seizures on wearables. Our first work [1] presents tiny machine learning for Arduino-based seizure detection. We train CNN models on melspectrogram-extracted EEG data, turning seizure classification into an image-based task.

PROS: An efficient Pattern-Driven Compressive Sensing Framework for Low-Power Biopotential-based Wearables with On-chip Intelligence.

ACM MobiCom 2022 (Acceptance rate: 17.8%)

Nhat Pham, Hong Jia, Minh Tran, Tuan Dinh, Nam Bui, Young Kwon, Dong Ma, Phuc Nguyen, Cecilia Mascolo, and Tam Vu

This study proposes PROS, an efficient pattern-driven compressive sensing framework for low-power biopotential-based wearables. PROS eliminates the conventional trade-off between signal quality, response time, and power consumption by introducing tiny pattern recognition primitives and a pattern-driven compressive sensing technique that exploits the sparsity of biosignals. Specifically, we (1) develop tiny machine learning models to eliminate irrelevant biosignal patterns, (2) efficiently perform compressive sampling of relevant biosignals with appropriate sparse wavelet domains, and (3) optimize hardware and OS operations to push processing efficiency.

IoTree: A Battery-free Wearable System with Biocompatible Sensors for Continuous Tree Health Monitoring

ACM MobiCom 2022 (Acceptance rate: 17.8%)

Tuan Dang, Trung Tran, Khang Nguyen, Tien Pham, Nhat Pham, Tam Vu, and Phuc Nguyen.

We present a low-maintenance, wind-powered, batteryfree, biocompatible, tree wearable, and intelligent sensing system, namely IoTree, to monitor water and nutrient levels inside a living tree. IoTree system includes tiny-size, biocompatible, and implantable sensors that continuously measure the impedance variations inside the living tree’s xylem, where water and nutrients are transported from the root to the upper parts. 

Detection of Microsleep Events with a Behind-the-ear Wearable System

IEEE Transactions on Mobile Computing 2023 (IF=7.9)

ACM MobiSys 2020 (Acceptance rate: 19.4%)

N. Pham, T. Dinh, Z. Raghebi, T. Kim, N. Bui, P. Nguyen, H. Truong, F. Banaei-Kashani, A. Halbower, T. Dinh, and T. Vu

WAKE is a novel wearable device that detects microsleep by monitoring biosignals from the brain, eye movements, facial muscle contractions, and sweat gland activities from behind the user’s ears. To ensure real-world reliability, WAKE introduces a Three-fold Cascaded Amplifying technique to tame motion artefacts and environmental noises. We evaluate the system with our custom-built prototype on 19 subjects and obtain 76% precision and 85% recall, showing the feasibility of microsleep detection on unseen subjects. 

DroneScale: Drone Load Estimation Via Remote Passive RF Sensing

ACM SenSys 2020 (Acceptance rate: 20.7%)

P. Nguyen, V. Kakaraparthi, N. Bui, N. Umamahesh, N. Pham, H. Truong, Y. Guddeti, D. Bharadia, E. Frew, R. Han, D. Massey, T. Vu.

DroneScale is a novel passive RF system that monitors the wireless signals transmitted by commercial drones and then confirms their models and loads. Its key technical contribution is a proposed technique to passively capture vibration at high resolution (i.e., 1Hz vibration) from afar, which was not possible before. 

Epileptic Seizure Detection and Experimental Treatment: A Review

Frontiers in Neurology 2020

Taeho Kim, Phuc Nguyen, Nhat Pham, Nam Bui, Hoang Truong, Sangtae Ha, Tam Vu

This article discusses recent advances in seizure sensing, signal processing, time- or frequency-domain analysis, and classification algorithms to detect and classify seizure stages. We also show a strong potential for applying recent advancements in non-invasive brain stimulation technology to treat seizures. 

Painometry - Wearable and objective quantification system for acute postoperative pain

ACM MobiSys 2020 (Acceptance rate: 19.4%)

H. Truong, N. Bui, Z. Raghebi, M. Ceko, N. Pham, P. Nguyen, A. Nguyen, T. Kim, et al.

Painometry is a wearable system for objective quantification of users’ pain perception based on multiple physiological signals and facial expressions. The lightweight form factor and a minimal number of sensors enable the mobility of Painometry and its capability as a daily wearable device. 

Earable Computing - An Ear-Worn Biosignal Sensing Platform 

ACM MobiSys 2019 Demo

Nhat Pham, Taeho Kim, Frederick M Thayer, Anh Nguyen, and Tam Vu.

Earable is a novel ear-worn biosensing platform for cognitive state quantification and human-computer interaction. Earable can capture biosignal, including brain waves activities, eyes movements, and facial muscle contractions from the back of the ears. Its form-factor is also convenient to use in everyday life.

eBP - Blood Pressure Measurement from inside the ear  

ACM MobiCom 2019 (Acceptance rate: 18.9%)

Best Paper Award.

ACM SIGMOBILE 2020, GetMobile 2019, CACM 2021 Research Highlights.

N. Bui, N. Pham, J. Barnitz, Z. Zou, P. Nguyen, H. Truong, T. Kim, N. Farrow, A. Nguyen, J. Xiao, R. Deterding, T. Dinh and T. Vu.

eBP is a novel wearable that can measure blood pressure from inside the user’s ear. It aims to minimize the measurement’s impact on users’ normal activities while maximizing its comfort. 

MSHCS-MAC: A MAC protocol for Multi-hop cognitive radio networks based on Slow Hopping and Cooperative Sensing approach

IEEE Symposium on Computers and Communications 2018 (Master's Thesis)

Nhat Pham, Kiwoong Kwon, and Daeyoung Kim.

MSHCS-MAC is a Cognitive Radio MAC protocol for multi-hop networks based on the Slow Hopping and Cooperative Sensing approach. MSHCS-MAC supports multi-hop communication without a dedicated control channel and multiple transceivers. It also integrates essential features such as bootstrapping, multi-channel operation, cooperative spectrum sensing, and time synchronization.

GS1 Global Smart Parking System: One Architecture to Unify Them All

IEEE International Conference on Services Computing 2017

Nhat Pham, Muhammad Hassan, Hoang Minh Nguyen and Daeyoung Kim.

This study aims to open the discussion to realize a global and common base for smart parking services by proposing a global system based on the GS1 standard. Our proposed architecture could be used globally and easily extended with different services by utilizing common and global standards.

Towards an Open Framework for Home Automation Development

IEEE ACOMP 2015 (Undergraduate Research)

Dang-Nhat Pham-Huu, Van-Hien Nguyen, Van-Anh Trinh, Van-Hieu Bui, and Hoang-Anh Pham.

This project proposes a framework for developing a home automation system based on an IPv6 personal area network protocol (6LoWPAN). The framework can automatically perform the appropriate actions based on user-predefined scenarios and provides runtime configurations without reprogramming or resetting the whole system.