Self-driving automotive technology took a big step forward this week, as NXP Semiconductors rolled out a pizza-box-sized module that could bring production-level, "high automation" to the roads by 2020.
Known as BlueBox, the new module contains the electronics for radar-, lidar-, and vision-based processing, as well as control of the autonomous steering, brakes, engine, and gearbox. The unit promises to take vehicles to Level 4 autonomy, in which the car operates itself but requires a driver to be present in the front seat. NXP said that four of the world's top five automakers are already developing vehicles on BlueBox.
Click on the image below to check out NXP's Bluebox tech:
"This is an architecture for production and mass deployment," Allan McAuslin, product line manager for vision and automated drive at NXP, told Design News .
The introduction is important because it brings the automotive world closer to the ultimate goal of full autonomy. The auto industry currently recognizes five levels of automated driving:
- Level 1 , which includes adaptive cruise control, lanekeeping, and automatic braking, is already on the roads in many newer vehicles.
- Level 2 , or partial automation, includes emergency braking with steering.
- Level 3 , known as conditional automation, adds chauffeuring capability for highway driving, but not for heavy traffic.
- Level 4, calls for a driver to be present in the front seat, but enables the vehicle to drive itself.
- Level 5 , which allows the "driver" to sit in rear seat, is the only higher level of automation.
NXP said that automakers currently deploying BlueBox are targeting Level 4, but not Level 5.
The announcement is also important for another reason -- it suggests that autonomy is moving beyond the realm of demonstration vehicles and into production in the near future. Design cycles for many 2020 models have already begun, NXP noted. "The automakers are not just talking about 2020 vehicles, they're actually designing them at this point," McAuslin said.
The BlueBox platform contains the processing capabilities to handle autonomous driving sensors, including radar, lidar, ultrasound, and cameras, as well as conventional motion and wheel speed sensors. It also incorporates vehicle-to-vehicle and vehicle-to-infrastructure communications, which would enable an autonomous car to "talk" with other vehicles, stop signs, and traffic lights.
BlueBox's computing platform contrasts sharply with predecessors that received notoriety during DARPA's famous Grand Challenge and Urban Challenge autonomous vehicles races between 2005 and 2007. Back then, self-driving vehicles were referred to as "supercomputers on wheels" because their computing hardware could fill a trunk or the entire back end of an SUV. BlueBox, in contrast, packs its computing prowess into metal box measuring about 14 inch x 18 inch x 2 inch. It contains two main