COMPASS
A system that provides real-time postural feedback, reducing driver fatigue and maintaining cognitive performance
Image Credit: Samuele Mazzei
Situation: Suboptimal driving posture is a primary contributor to muscular fatigue and cognitive decline, yet current automotive research lacks real-time, closed-loop interventions to actively break this cycle.
Task: Design and validate COMPASS (Corrective Multisensory Posture Assistance System for Seating), a portable, multisensory system to proactively detect and correct poor posture, thereby enhancing driver safety and alertness.
Action: Utilize sensor fusion (NIR camera and pressure mat) to track torso kinematics and weight distribution. Employ Machine Learning (CNN/Random Forest) to define ergonomic baselines, using sEMG physiological data for ground truthing. Design pilot testing to refine feedback modalities and a within-subject experimental design for ecological validation in real-world driving.
Result: Investigated a system that significantly reduces muscle strain and maintains cognitive performance (vigilance scores). Delivered an open-source model designed for easy retrofitting into commercial fleets.
This project was just a research proposal and has not been implemented in real life yet.