Radar processors aim for Level 4 autonomy

NXP Semiconductors has introduced its latest S32R47 imaging radar processors, utilizing 16 nm FinFET technology. This third generation of imaging radar processors boasts double the processing power of its predecessor, while also enhancing system cost and power efficiency. When paired with mmWave radar transceivers, power management solutions, and in-vehicle networking products from NXP, the S32R47 family meets the functional safety ASIL ISO 26262 ASIL B(D) requirements, positioning the automotive industry for advancements in autonomous driving.

As projected by Yole Intelligence’s Status of the Radar Industry 2024 report, by 2029, an estimated 40% of vehicles on the road will be passenger cars equipped with driving automation Level 2+ (L2+) and Level 3 (L3), alongside a rising number of vehicles featuring Level 4 (L4) automation. To cater to the evolving demands of the expanding autonomous driving market for SDVs, automotive OEMs and tier 1 suppliers must enhance radar performance. This enhancement is crucial for the safe implementation of advanced autonomy features like piloted driving and fully automated parking.

Meindert van den Beld, Senior Vice President & General Manager of Radar and ADAS at NXP, highlighted the capabilities of the S32R47, stating that it can efficiently process three times or more antenna channels in real-time compared to current offerings. This advancement enables improved imaging radar resolution, sensitivity, and dynamic range, essential for demanding autonomous driving scenarios, while still meeting the strict power and cost requirements set by OEMs for mass production.

Imaging radar technology leverages richer point cloud data to create detailed environmental models, crucial for AI-based perception systems that support assisted and autonomous driving in complex urban settings. The S32R47 is equipped with a high-performance multi-core radar processing system, enabling denser point cloud output and advanced algorithms that bolster next-generation ADAS systems. These enhancements result in better object separability, heightened detection reliability, and more precise classification of entities like vulnerable road users or misplaced cargo.

Developed on the foundation of two previous generations, the S32R47 delivers up to 2x processing performance in the radar MPU within a 38% smaller IC footprint. Additionally, it incorporates AI/ML support for features such as enhanced Direction of Arrival (DoA) processing and object classification. The S32R47 achieves comparable or superior performance with up to 89% fewer antenna channels than alternative products, addressing integration challenges while reducing system cost, size, and power consumption.