Advancing Intoxication Detection: A Smartwatch-Based Approach

TL;DR

This work tackles non-invasive intoxication detection via smartwatch sensors to enable just-in-time adaptive interventions. It presents a three-week, multi-sensor dataset (accelerometer, gyroscope, heart rate, TAC) from 30 participants, labels intoxication using TAC thresholds around $35~\mu g/L$ (≈ BAC $0.05\%$), and evaluates seven models including HDC and 1D-CNN for time-series classification. Results show that HDC and 1D-CNN offer the best balance between accuracy and efficiency for edge deployment, with feasible inference on mobile hardware demonstrated via Executorch on an Android device. The study advances IoT-enabled JITAIs by providing a practical, energy-aware approach to real-time intoxication warnings and potential interventions, supporting healthier drinking habits in real-world settings.

Abstract

Excess alcohol consumption leads to serious health risks and severe consequences for both individuals and their communities. To advocate for healthier drinking habits, we introduce a groundbreaking mobile smartwatch application approach to just-in-time interventions for intoxication warnings. In this work, we have created a dataset gathering TAC, accelerometer, gyroscope, and heart rate data from the participants during a period of three weeks. This is the first study to combine accelerometer, gyroscope, and heart rate smartwatch data collected over an extended monitoring period to classify intoxication levels. Previous research had used limited smartphone motion data and conventional machine learning (ML) algorithms to classify heavy drinking episodes; in this work, we use smartwatch data and perform a thorough evaluation of different state-of-the-art classifiers such as the Transformer, Bidirectional Long Short-Term Memory (bi-LSTM), Gated Recurrent Unit (GRU), One-Dimensional Convolutional Neural Networks (1D-CNN), and Hyperdimensional Computing (HDC). We have compared performance metrics for the algorithms and assessed their efficiency on resource-constrained environments like mobile hardware. The HDC model achieved the best balance between accuracy and efficiency, demonstrating its practicality for smartwatch-based applications.

Figures (4)

• Figure 1: Example user session with TAC spikes above 35$\mu$g/L
• Figure 2: Mobile application screens.
• Figure 3: Smartwatch JITAI intervention.
• Figure 4: Clustering of Users based on TAC value