BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

NXP Launches New Imaging Radar Chips For ADAS And Automated Driving

Following
This article is more than 2 years old.

2022 is expected to be a big year for new sensor types to be added to production vehicles. Cameras, ultrasonic sensors and basic low-resolution radar sensors are already commonplace on even mainstream vehicles. This year a number of new models will hit the streets with lidar sensors and high-resolution imaging radar is also expected to play a key role in improving the capabilities of advanced driver assist systems (ADAS) and automated driving systems (ADS). Dutch chipmaker NXP hopes to be a key player in this market. 

NXP already produces the radar processing chips used in sensors from a number of suppliers and hopes to grow that market with the new S32R45 and S32R41 chip families. The S32R45 is targeted at L4 and above ADS applications while the S32R41 is optimized for hands-free partially automated ADAS applications. Both chips are based on a mix of ARM Cortex A53 and M7 cores that provide enough performance to do the necessary signal processing.  

Current ADAS radar sensors that are used for adaptive cruise control and blindspot monitoring systems typically have about 6 channels and measure speed and distance of objects in a single plane. This limits the ability to distinguish between different objects or for example vehicles on the road vs an overpass. This is why these sensors are designed to focus on a specific region of interest. 

For example, a long-range forward facing sensor is looking mainly straight ahead and ignoring objects off to the sides. A single sensor of this type is inadequate for ADS and will have difficulty distinguishing a vehicle stopped on the side of a road such as an emergency vehicle. Tesla’s TSLA with Autopilot have had numerous crashes using this type of radar. However, rather than upgrading to better sensors, Tesla has opted to drop the radar and rely on cameras only. 


However, while camera images are very useful for classifying what different objects in the field of view are, cameras can have difficulty seeing in harsh lighting, poor weather conditions or at night. Depending on how cameras are configured on a vehicle they may not be suitable for accurate distance and speed measurements unless multiple cameras are facing the same direction. Radio frequency signals from radar can provide very accurate speed and distance measurements even through atmospheric obstructions such as fog, snow and heavy rain. 

The NXP S32R45 chips can enable radar with up to 4 radar transceivers and 192 virtual channels with coverage across both horizontal and vertical directions. The S32R41 can support up to 2 transceivers and 48 virtual channels. This enables the generation of something that looks more like a lidar point cloud so that numerous objects can be tracked independently in multiple planes. The S32R chips can also enable time division multiplex control of the transceiver for multi-mode imaging. 

This allows a single sensor to provide long, mid and short-range detection at different fields of view all concurrently. NXP powered sensors can achieve angular resolution of less than 1 degree. The chips also incorporate a linear algebra accelerator core to help speed up perception capabilities. 

Production of the S32R45 began in late 2021 and the first applications should arrive sometime in 2022. The S32R41 should debut in 2023.

Follow me on LinkedInCheck out my website