Kinematic signatures of impairment: Detecting alcohol intoxication in e-scooter riders using sensor data and machine learning

Rahul Rajendra Pai; Marco Dozza; Alexander Rasch; Ali Mohammadi; Marco Capuccini

cs.LGSep 29, 2026

Kinematic signatures of impairment: Detecting alcohol intoxication in e-scooter riders using sensor data and machine learning

Authors: Rahul Rajendra Pai, Marco Dozza, Alexander Rasch, Ali Mohammadi, Marco Capuccini

Organizations: Chalmers University of Technology, Division of Vehicle Safety, Department of Mechanical Engineering, Gothenburg, Sweden · Voi Technology AB, Stockholm, Sweden

Abstract

Alcohol intoxication is a leading contributor to fatal and severe-injured e-scooterist crashes. Current countermeasures, such as temporal restrictions or pre-ride cognitive screening, cannot continuously assess an e-scooterist's physical motor control or impairment in real time. We conducted a controlled experiment in which 25 participants rode an instrumented e-scooter through a test track while sober and at two targeted blood alcohol concentration levels (0.05% and 0.08%). The e-scooter was instrumented with a six-axis inertial measurement unit (IMU), and throttle and brake lever position sensors, all sampled at 100 Hz. Two complementary signal features were computed: normalised permutation entropy, which quantifies temporal complexity, and standard deviation, which quantifies signal amplitude. Repeated measures correlation identified seven kinematic features (all IMU and throttle signals) whose entropy decreased (p < 0.001) while standard deviation increased (p < 0.01) with increasing intoxication, indicating that intoxicated riders shift from continuous, low-amplitude micro-corrections to fewer, high-amplitude reactive corrections. An entropy based multi-class logistic regression classifier, evaluated through leave-one-participant-out cross-validation, achieved 85% overall accuracy and a weighted one-vs-rest area under the receiver operating characteristic curve (AuROC) of 0.94, with a sober-vs-high AuROC of 1.00. Steering rate and lateral acceleration were the most important predictive features, indicating that alcohol induces a distinct collapse in lateral equilibrium during riding. Ultimately, these results demonstrate that onboard kinematic sensing combined with entropy-based signal analysis can reliably distinguish sober from intoxicated e-scooter riding, providing a foundation for automatic intoxication detection systems that preserve mobility for sober riders.

Figures & tables

Fig. 1: Instrumentation on the vehicle (Panel A) and the layout of the MicroLab test track (Panel B).

Feature	PE $p$ -value	SD $p$ -value
$r_{y}$ (steering rate)	$2.50\times 10^{-28}$	$1.75\times 10^{-10}$
$r_{z}$ (roll rate)	$1.79\times 10^{-24}$	$7.81\times 10^{-13}$
$a_{x}$ (lateral acceleration)	$5.75\times 10^{-21}$	$8.68\times 10^{-3}$
$r_{x}$ (pitch rate)	$1.01\times 10^{-16}$	$1.07\times 10^{-14}$
throttle	$1.92\times 10^{-8}$	$9.07\times 10^{-3}$
$a_{y}$ (vertical acceleration)	$6.47\times 10^{-6}$	$1.98\times 10^{-8}$

Table 1: Repeated measures correlations between features and intoxication level. PE was tested with

H_{1}\colon\rho<0

(complexity decreases); SD was tested with

H_{1}\colon\rho>0

(amplitude increases).

Fig. 2: Within-subject centred (WSC) permutation entropy (PE) distributions for all nine evaluated signals stratified by intoxication condition (sober, low, high).

Fig. 3: Within-subject centred (WSC) standard deviation (SD) distributions for all nine evaluated signals stratified by intoxication condition (sober, low, high).

		High			Low			Sober
Model	Accuracy	Precision	Recall	F1 score	Precision	Recall	F1 score	Precision	Recall	F1 score
Sum entropy	0.74	0.78	0.78	0.78	0.62	0.68	0.65	0.85	0.76	0.80
WSC-PE logistic regression	0.85	0.87	0.89	0.88	0.81	0.76	0.78	0.87	0.91	0.89
Support vector machine	0.86	0.86	0.96	0.91	0.84	0.74	0.79	0.87	0.89	0.88
WSC-SD logistic regression	0.61	0.62	0.63	0.62	0.53	0.50	0.52	0.68	0.71	0.70

Table 2: Three-class classification performance (LOPO CV, 25 folds).

Model	Sober-vs-rest	Low-vs-rest	High-vs-rest	Weighted	Sober-vs-high
Sum entropy	0.9475	0.7923	0.9297	0.8867	0.9937
WSC-PE logistic regression	0.9664	0.8912	0.9682	0.9403	0.9995
Support vector machine	0.9667	0.9002	0.9625	0.9417	0.9884
WSC-SD logistic regression	0.8782	0.6936	0.8206	0.7940	0.9024

Table 3: AuROC scores under LOPO CV.

Fig. 4: Mean standardised entropy-based logistic regression coefficients for the seven retained features across the three intoxication classes.

Feature	$\Sigma\,\|\beta\|$
$r_{y}$ (steering rate)	3.21
$a_{x}$ (lateral acceleration)	3.03
$r_{z}$ (roll rate)	1.93
throttle	1.76
$r_{x}$ (pitch rate)	1.47
$a_{z}$ (longitudinal acceleration)	0.60

Table 4: Feature importance ranking based on the summed absolute mean WSC-PE-based LR coefficients across all LOPO folds.

Explore similar work

May 22, 2026cs.HC

Detecting Drunk Driving Using Off-the-Shelf Smartwatches

Alcohol-impaired driving remains a major yet preventable cause of road traffic injury and death, with many drivers underestimating their level of intoxication. Compared to in-vehicle systems, mobile drunk-driving detection using consumer smartwatches offers a scalable way to trigger preventive interventions and increase awareness without additional in-vehicle hardware. We introduce a system that leverages wrist accelerometer data and heart rate variability-derived physiological signals to detect alcohol-related driving impairment. We collected data in a randomized, controlled three-arm test-track study (n=54) and trained both logistic regression models with window-aggregated features and a two-tower 1D convolutional neural network (CNN), to detect alcohol-impaired driving. The CNN achieved a participant-averaged area under the receiver operating characteristic (AUROC) of 0.88 for detecting any alcohol intoxication and 0.86 for detecting driving above the WHO-recommended limit of 0.05 g/dL. To the best of our knowledge, this is the first work to (1) demonstrate drunk-driving detection using consumer smartwatches, (2) develop and evaluate such a system in a real vehicle on a closed test track, and (3) rigorously assess generalization to unseen participants. Together, these findings highlight the potential of wearable-based sensing to support scalable, measurement-driven prevention of alcohol-related traffic harm.

Robin Deuber, Lanlan Yang, Michal Bechny +9

ETH Zürich, Switzerland · University of Zürich, Switzerland · University of Bern, Switzerland +2

Accelerometer Data

Ask PDF

Jul 4, 2026cs.CV

EgoInertia-MI: A Multimodal Egocentric Vision and IMU Benchmark for Motor Impairment Assessment

Motor impairments, including tremor, bradykinesia, gait abnormalities, and postural instability, are common across many neurological and movement-related conditions. Conventional clinical assessments are often intermittent and may fail to capture subtle temporal variations in motor behavior. While wearable IMUs and third-person video have shown promise for objective motor assessment, third-person recordings raise privacy concerns and require constrained acquisition setups. In contrast, egocentric vision provides a more naturalistic and privacyaware alternative. In this work, we introduce EgoInertia-MI, a multimodal benchmark dataset combining synchronized egocentric video and wearable IMU signals for motor impairment analysis. The dataset contains 19 upper- and lower-body activities performed by healthy volunteers simulating varying levels of motor impairment severity levels: no impairment, mild impairment, and severe impairment. We establish two benchmark tasks: action recognition and motor impairment severity estimation, and evaluate multiple unimodal and multimodal baselines. Experimental results show that egocentric video provides strong cues for motor impairment assessment, while multimodal fusion achieves the best overall performance, reaching 0.78 Macro-F1 for severity estimation and 0.93 Macro-F1 for action recognition. These findings highlight the potential of combining egocentric vision and wearable sensing for ecologically valid and privacy-aware motor assessment. Code and data are available at:https://fatemah-alh.github.io/EgoInertia-MI-Page/.

Fatemah Alhamdoosh, Pietro Pala, Abduallah Mohamed +1

University of Florence, Italy · Meta Reality Labs, USA · University of Edinburgh, UK

Inertial Measurement Units Egocentric Vision Gait Analysis Egocentric Video Inertial

Ask PDF

May 21, 2026cs.CV

MOTOR: A Multimodal Dataset for Two-Wheeler Rider Behavior Understanding

Two-wheelers account for a disproportionately high share of road fatalities in the Global South. Research on two-wheeler rider behavior, however, lags far behind four-wheelers, where multimodal datasets have driven major advances in Advanced Driver Assistance Systems (ADAS). To address this gap, we present the MOtorized TwO-wheeler Rider (MOTOR) dataset, the first large-scale, multi-view, multimodal resource dedicated to two-wheelers in dense, unstructured traffic. MOTOR comprises 1,629 sequences (25+ hours of video data) collected from 16 riders and integrates synchronized front, rear, and helmet videos, rider eye-gaze from wearable trackers, on-road audio, and telemetry (GPS, accelerometer, gyroscope). Rich annotations capture traffic context, rider state, 12 riding maneuvers spanning conventional and unconventional behaviors, and legality labels (Legal, Illegal, Unspecified). We benchmark rider behavior recognition and maneuver legality classification using state-of-the-art video action recognition backbones (CNN and Transformer-based), extended with multimodal fusion, and find that combining RGB, gaze, and telemetry consistently yields the best performance. MOTOR thus provides a unique foundation for advancing safety-critical understanding of two-wheeler riding. It offers the research community a benchmark to develop and evaluate models for behavior analysis, legality-aware prediction, and intelligent transportation systems. Dataset and code is available at https: //varuniiith.github.io/MOTOR-Dataset/

Varun A. Paturkar, Shankar Gangisetty, C. V. Jawahar

CVIT, IIIT-Hyderabad, India

Naturalistic Driving Data Multimodal Dataset Human Activity Recognition Agent Behavior Modeling

Ask PDF