Google's Self-Driving Cars Are Finally Taking to the Freeway
The Alphabet subsidiary, now known as Waymo, has announced that its robotaxi service will now be able to drive on freeways in several cities across the US and abroad. This marks a significant milestone for the company's self-driving car project, which began testing in the Bay Area back in 2009.
To prepare for highways, Waymo engineers supplemented real-world driving data with data collected on private courses and simulations. Two onboard computers create system "redundancies," meaning vehicles will have computer backup systems if something goes wrong. The vehicles are also trained to exit highways in emergency situations but can pull over as well.
According to Naomi Guthrie, a Waymo user experience researcher, freeway driving is one of those things that's easy to learn but hard to master. Highways are predictable with clear signs and lane lines, limited vehicles and players, making them easier for the vehicle's software to recognize and predict.
However, Waymo executives say safety emergencies on highways are rare, so the team was unable to collect enough real-world data to train its vehicles to operate safely there. Engineers supplemented this data by creating simulations and using private courses with closed courses. They have also worked closely with law enforcement agencies to develop procedures for stranded vehicles on highway shoulders.
At launch, Waymo's freeway service will only be available in limited portions of certain highways across the US and abroad, but it plans to roll out the feature to all riders eventually.
The Alphabet subsidiary, now known as Waymo, has announced that its robotaxi service will now be able to drive on freeways in several cities across the US and abroad. This marks a significant milestone for the company's self-driving car project, which began testing in the Bay Area back in 2009.
To prepare for highways, Waymo engineers supplemented real-world driving data with data collected on private courses and simulations. Two onboard computers create system "redundancies," meaning vehicles will have computer backup systems if something goes wrong. The vehicles are also trained to exit highways in emergency situations but can pull over as well.
According to Naomi Guthrie, a Waymo user experience researcher, freeway driving is one of those things that's easy to learn but hard to master. Highways are predictable with clear signs and lane lines, limited vehicles and players, making them easier for the vehicle's software to recognize and predict.
However, Waymo executives say safety emergencies on highways are rare, so the team was unable to collect enough real-world data to train its vehicles to operate safely there. Engineers supplemented this data by creating simulations and using private courses with closed courses. They have also worked closely with law enforcement agencies to develop procedures for stranded vehicles on highway shoulders.
At launch, Waymo's freeway service will only be available in limited portions of certain highways across the US and abroad, but it plans to roll out the feature to all riders eventually.