What's on your mind? A Mental and Perceptual Load Estimation Framework towards Adaptive In-vehicle Interaction while Driving
URL: http://arxiv.org/abs/2208.05564v1
archive: archived pipeline: cataloged verified
Abstract
Several researchers have focused on studying driver cognitive behavior and mental load for in-vehicle interaction while driving. Adaptive interfaces that vary with mental and perceptual load levels could help in reducing accidents and enhancing the driver experience. In this paper, we analyze the effects of mental workload and perceptual load on psychophysiological dimensions and provide a machine learning-based framework for mental and perceptual load estimation in a dual task scenario for in-vehicle interaction (https://github.com/amrgomaaelhady/MWL-PL-estimator). We use off-the-shelf non-intrusive sensors that can be easily integrated into the vehicle's system. Our statistical analysis shows that while mental workload influences some psychophysiological dimensions, perceptual load shows little effect. Furthermore, we classify the mental and perceptual load levels through the fusion of these measurements, moving towards a real-time adaptive in-vehicle interface that is personalized to user behavior and driving conditions. We report up to 89% mental workload classification accuracy and provide a real-time minimally-intrusive solution.
Summary
Dual-task driving study (in-vehicle simulator) developing a machine-learning framework that separately estimates mental workload (MWL) and Lavie's perceptual load (PL) from off-the-shelf non-intrusive psychophysiological sensors. Statistical analyses show MWL drives clear changes across psychophysiological dimensions while PL effects are small. A classifier fusing the sensor channels reaches up to 89% MWL classification accuracy, supporting a real-time minimally-intrusive adaptive in-vehicle interface that can personalize to driver state.
Key finding
Mental workload and perceptual load are dissociable from psychophysiological signals during driving (MWL strongly, PL weakly), and a fused-sensor ML classifier achieves up to 89% MWL classification accuracy suitable for adaptive in-vehicle HMIs.
Methodology
Exp 1: 10 participants, repeated measures across 6 sessions from 26 total. Exp 2: 20 participants, Old/New sequence comparison. On-road driving paradigm with DRT and NASA-TLX measures.
Sample size: Exp 1: N=10; Exp 2: N=20
Quality score: 5 / 5