Investigators at the AgeLab continue to make headway into the expanding field of driving behavior research with three presentations at the New England Chapter of the Human Factors and Ergonomics Society’s Annual Student Conference earlier this month. Most notably, MIT AgeLab Visiting student Mauricio Muñoz won the Volpe Award for Best Transportation Presentation for his project “Using Hidden Markov Models for Classification of Driver Visual Behavior”.
Muñoz’s work investigated the relationship between drivers’ eye glance patterns and head rotation while driving. He applied advanced statistical analyses via multiple machine learning methods in order to predict the location of eye glances based upon a drivers head position. Muñoz found that Hidden Markov Models outperformed other algorithms to produce a high level of predictive accuracy. By using this statistical method, results showed that head-rotation variables could be used as a strong indicator of glance-location predictions inside of cars. This is of interest since relatively low cost camera-based technologies for monitoring head position are becoming available, which could be used to alert the driver when they have become distracted and need to direct their attention back to the forward roadway.
The other two presentations touched on heightened concerns about increased driver attentional demand due to the presence of new technology inside of vehicles. “The Impact of Secondary Task Modality on Field Driving Performance” by Sinelnikova and colleagues used predictive models to examine differences in driving performance while drivers tuned the radio vocally and manually. Her work suggests that the modality used to interact with the radio interface may be a metric that is highly predictable based upon measures of longitudinal and lateral vehicle control.
“Assessing the Relative Impact of Smartwatch and Smartphone Use on Workload, Attention, and Driving Performance in a Driving Simulator” by Domel and colleagues investigated the effects of smartwatches on driving performance. As a relatively new technology for which only limited behavioral research currently exists, smartwatches are sure to infiltrate the driver’s seat in the upcoming months and years. Domel et al.’s project compared the workload associated with placing a phone call from the smartwatch (vocally) versus a traditional smartphone (both vocally and manually). They found that vocal dialing, regardless of device, resulted in less erratic driving behavior than manual dialing.