NXP Semiconductors recently shrank the gap between automotive and computer world to some extent by bringing forth an open computing platform “BlueBox”. The main function of this platform is to support automakers in designing and testing the potential of self-driving cars. Basically, BlueBox is a centralized computer engine that abstracts energy from bits flowing in from sensors that are spread across the vehicle. It forms a network between all sensors fitted across different parts of an autonomous vehicle including the Lidar, cameras, and radar systems. The information collected by these sensors is routed to the BlueBox for processing and formulating a 3D model of vehicle’s surrounding ambience.
There is no second opinion to the fact that systems that can analyze sensor data in real-time have vast capability to improve decision making ability of autonomous vehicles. Whether they need to slow down in upcoming traffic or slam the breaks, the data relayed from sensors can be the best source for making quick decisions. BlueBox can certainly play a vital role in accomplishing this task. It can easily enhance the potential of driver assisted systems that will alert him under dangerous circumstances and give the system to carry out the required actions like applying emergency brakes or give the blind-spot warnings.
The engine developed by NXP has several other modified features that will give the in-fitted car computer better control over car’s systems. As the company likes to quote, “the embedded intelligence and machine learning required for complete situational assessments, supporting advanced classification tasks, object detection, localization mapping and vehicle driving lessons.” The platform also comes with an overboard system that will share the positional data of the car with other vehicles on the road.
The core of the system bears two automotive processors. The first part is the S32V automotive vision processor, a Freescale holdout. The system is capable enough of receiving images from cameras installed around the car, extract and identify relevant objects, and formulate a 3D models. However, you will also need a graphics processor unit to complete this task. While the first processor comes from Cognivue, the second half of BlueBox gets added from NXP’s embedded catalog. The platform consists of an embedded computer processor that comes with 64-bit ARM core that count to eight in total. These empower BlueBox to pass 90,000 million instructions per second. The company will display the system at the NXP FTF Technology Forum taking place in Austin this week.
Even before it is being commercialized, BlueBox is already under scrutiny of five world level automobile manufacturing companies. BlueBox can prove to be a milestone not only in development of innovative safety systems but also take automakers a step closer to the dream of self-driving cars in next few years.