NXP Unveils Platform for Autonomous Vehicles

At its FTF show this week, the chip maker is demonstrating its offering, which includes its BlueBox compute engine for self-driving cars.

NXP car platform

Officials with chip maker NXP Semiconductors expect the company to be a significant player in the fast-growing autonomous vehicle market with a new platform they're rolling out that includes everything from the sensors that collect massive amounts of data to the compute engine capable of fusing all that data and ensuring that the right decisions are made.

At the company's FTF Technology Conference this week in Austin, Texas, NXP officials will demonstrate its platform in a 3D printed car. The introduction of a platform rather than a point product is a key differentiator in a market that is garnering a lot of attention from other chip makers and technology vendors, according to Matt Johnson, vice president and general manager for automotive microcontrollers and processors at NXP.

"Everyone is talking about one product or one app," Johnson told eWEEK. "This is the entire platform."

NXP, through its $12 billion acquisition last year of Freescale, has become a major tech supplier to the automotive market, having shipped more than 30 million advanced driver assistance system (ADAS) processors worldwide to date. In addition, eight of the world's 10 largest carmakers use ADAS processors from NXP.

Almost every point in the platform—including the BlueBox compute engine—is powered by NXP silicon, he said. That includes the sensors for everything from speed and motion to radar, Lidar and cameras, as well as the BlueBox and instrument panel and vehicle control aspects, such as steering, brakes and airbags.

What NXP won't supply is the bulk of the software. Because it's an open, Linux-based platform, developing applications for the platform is made easier, and OEMs and automakers will be able to quickly customize it for their needs, according to company officials. The foundation of NXP's offering is the BlueBox compute engine, which includes NXP's ARM-based 64-bit LS2088A embedded computer chip and its S32V automotive vision processor, which includes a graphics engine, high-performance image processing accelerators, and sensor fusion functionality.

The sensors on autonomous vehicles are used to track multiple inputs, and the data is then sent to the BlueBox, where it is brought together and fused to create a complete 360-degree model of the world around the car. The combination of the two NXP processors with embedded intelligence and machine learning capabilities enables everything from advanced classification and object detection to localization, mapping and vehicle driving decisions, officials said.

The BlueBox enables a performance of up to 90,000 DMIPS (million instructions per second) while consuming less than 40 watts of power.

According to NXP's Johnson, a key point of the platform is that it is available now. All the processors used in it are either shipping or are sampling and are ready for production, and four of the world's five largest carmakers already have the BlueBox engine now. It has been shipping to select customers since September 2015 and is available now to all customers.

"This market is a little bit of the wild west right now," he said. "Everybody is making big claims and headlines with what they're doing. … Everything we're talking about here [with the platform] is available to customers in real time."

When NXP bought Freescale last year, officials said the acquisition would help the company in a range of areas, including the Internet of things (IoT) and connected cars.

"The car industry is clearly changing in the next five to 10 years, much more than it's changed in the last 50 years," Kurt Sievers, executive vice president and general manager of NXP's Automotive Business, said after the deal closed in December 2015. "In combination, we are a powerhouse in the automotive industry. … We will make cars sense, think and act."

Johnson said the company is bringing a lot of experience to the autonomous vehicle space.

"We're not just dipping our toe here for the first time in this space," he said. "It's important to us, and we're hyper-focused on it."

The market is expected to grow quickly. Analysts with ABI Research last month said that the ADAS market itself will grow from $11 billion this year to $132 billion by 2026. And carmakers are planning to roll out self-driving cars within the next several years. BMW officials last week said they expected to have their first self-driving car—an electric vehicle called the i Next—on the road by 2021.