MADISON, Wis.—When NXP Semiconductors today (May 16) unveils at its own conference a 3D-printed car with a new autonomous engine — dubbed Bluebox — it will offer a glimpse of the future in which autonomous cars are developed on a modular and open vehicles platform, rather than inside a proprietary shell.
The growing complexity of the software and hardware required to develop autonomous cars is a force that threatens to break open the automotive industry’s long-established, closed-development environment.
Central to NXP’s comprehensive autonomous vehicles platform is its new Bluebox computing engine, capable of fusing data that’s processed from different sensor nodes such as radar, lidar, vision sensing and an onboard secure V2X system.
Calling it “an open platform based on Linux,” Matt Johnson, NXP’s vice president and general manager for automotive microcontrollers and processors, said, “It’s ready for OEMs and Tier Ones to develop and experiment their own autonomous cars.” The open platform allows carmakers to “develop their own unique software and algorithms on it,” he noted. Further, car OEMs can plug in other vendors’ sensor node processors, when necessary, he added. “It's unrealistic to think carmakers will design their cars using chips all coming from NXP.”
Ready for production
Perhaps most important, as Johnson explained, this is a platform “car OEMs can take to production.”
Noting that it’s “complete with auto-grade reliability, security and safety functionalities,” he touted a scalable platform on which carmakers can “develop their own, mainstream vehicles of Level 2, Level 3, and Level 4 autonomous cars for the commercial market.” Johnson stressed, “This [autonomous vehicles platform] is not a science experiment.”
The automotive analyst community has generally given a thumbs-up on NXP’s new Bluebox-based platform, citing its openness to OEMs, sensor-fusion capabilities, functional safety and reliability.
But analysts have questioned NXP’s readiness for the growing challenges of artificial intelligence (AI)-based machine vision and database mapping, which they regard as essential to truly replacing the driver in an autonomous car.