With confirmation that the Model 3 will include an interior camera that will likely be used to monitor driver behavior, it’s a good time to revisit Dr. Lex Fridman’s presentation at TMC Connect 2016.
He told the TMC audience that driver-facing cameras are a “huge missing piece” for advancing automation.
Fridman is a postdoctoral associate working on Deep Learning and Computer Vision in the context of Shared Autonomy and Intelligent Vehicle Systems at the Massachusetts Institute of Technology (MIT).
In this TMC Connect 2016 conference session, Fridman explains how his group uses computer vision and deep learning approaches to extract essential information about the car, the driver, and the external world during Tesla Autopilot use.
The group at MIT has conducted research on advanced vehicle technology use in naturalistic driving. At the time of his presentation, the group had outfitted eight Tesla Model S vehicles with cameras and other sensors capturing not only the external world, but driver state and vehicle state as well. Their goal was to expand the fleet of instrumented Tesla vehicles to record and analyze driver interaction with Autopilot over thousands of miles. It seems Tesla may soon have the ability to gather the same data from a fleet of thousands.