The Human Limits of Autonomous Cars

My recent road trip to Chicago was my first in a vehicle that had the complete current suite of driver assistance technologies, namely adaptive cruise control with autonomous braking, lane departure warning with lane keep assist, and blind spot monitoring. To be honest, it make the trip significantly less stressful and much more pleasant.

I had the adaptive cruise control on for the entire trip, with it typically set to give me four seconds of spacing between me and the vehicle in front of me. At full highway speeds, it seems that the four seconds is the right amount of spacing to give maximum flexibility for emergency maneuvering.

We’ve been talking a lot about driverless cars, but my experience illustrated the limits of automation in cars. Frequently drivers would see the space between me and the car in front of me and move over. The systems adapted in a less fluid way. While in manual mode, I would simply lay off the acceleration. Instead, the system stuck to its spacing requirement and braked, which may be unexpected for the driver behind me. Driver assistance systems always err on the side of safety, but human drivers to not.

When a car in front of me moved out of the way, the cruise control was also fairly slow on the acceleration. I could tell this also caused consternation with some drivers behind me, some of whom zoomed around to pass on the right. Slower acceleration is clearly preferable from a mechanical and fuel economy perspective, but humans are seldom rational when behind the wheel.

I’ve concluded that autonomous systems aren’t fully compatible with manual drivers. They may be excellent at forward collision mitigation, but I can easily foresee an increase in rear collisions by drivers who aren’t anticipating its abundant caution.

© 2019 Chet J. Collins