Self-driving cars, as popularised by the likes of Google and Tesla, are meant to be transportation bubbles that operate free of any human interference. But full autonomy isn’t the only option, and Toyota is investing in a system that would use computers as an aid to human drivers, not a replacement.
As MIT Technology Review reports, Toyota is investing in a new facility at the University of Michigan, as part of an effort to develop a new “guardian angel”. This would be a system that would temporarily wrest control from human drivers in order to avoid a collision.
It’s far from a novel idea: driver aids like traction control and anti-lock braking have been modifying a driver’s input for decades, and emergency automatic braking is going to be standard on pretty much every car by 2022. Toyota’s system would presumably go beyond that, telling the car to swerve (or not!), accelerate or brake as needed.
Toyota’s proposed future is far less exciting and futuristic than Google’s cute robotic pods, but it might also be more realistic, especially in the short term. Fully autonomous vehicles still have a lot of hurdles to cross—who’s going to be legally liable in a crash, for starters, and the fact that most autonomous cars suck when it snows.
Apart from the limitations on vehicles, it’s also worth remembering that our infrastructure is designed for human cars, not robotic drivers. In order to make the most of driverless technology, everything from intersections to speeds limits and crosswalks will have to be remade, and that’s not a small task.
Bearing all that in mind, Toyota’s plan seems like a sensible transition. Semi-autonomous cars will start getting people used to the idea of a computer driving them around. It will also allowing sensor-laden cars to drive billions of much-needed miles in the real world, gathering data that can make driverless AI more intelligent.[MIT Technology Review]