An exploratory study investigating users’ understanding of noise model uncertainties utilising webcam eye tracking

We report on an exploratory study focusing on how people interpret uncertainties in noise models related to road traffic, as assessed using the Common Noise Assessment Methods in Europe (CNOSSOS-EU). Specifically, via an online eye movement study with 35 participants, we investigate how viewers’ visual attention and behaviour can reveal uncertainties in studied uncertainty models. As a case study, we generated a preliminary noise model for Munich using the library and model builder by the NoiseModelling project. For simplicity, the examined model only accounts for road traffic noise and does not represent dynamic variations in noise levels throughout the day. Participants (n=35) engage in tasks using different noise maps and colour schemes, including those from the NoiseModelling documentation and ColorBrewer. The eye tracking data reveals significant patterns in user responses, including awareness of noise in major intersections, train stations, and residential areas. The study also assesses the performance of the participants while using RealEye.io’s webcam-based eye tracking across devices: desktops, tablets, and smartphones. The participants using desktops exhibit the highest performance, while participants using smartphones show the lowest. Our exploratory study reveals the importance of device-specific considerations in eye tracking-based cartographic user studies and suggests future work to tailor stimuli for each device type.