Supports AI processing required for automotive front-view cameras with industry-leading ultra-low-power CNN IP

 Tokyo, Japan, February 28, 2018 – With the rapid advancement of advanced driver assistance systems (ADAS) and autonomous driving technology, autonomous vehicles are becoming a reality. Renesas Electronics Corporation (TSE: 6723), the world's leading supplier of automotive semiconductor solutions, today announced the launch of the new R-Car V3H system-on-chip (SoC). The SoC provides powerful computer vision performance and AI processing capabilities for automotive front-view cameras with industry-leading low power consumption, and is suitable for mass-produced Level 3 (conditional automation) and Level 4 (highly automated) autonomous vehicles. Optimized for stereo front-view camera applications, the new R-Car V3H SoC delivers five times the computer vision performance of the R-Car V3M SoC for NCAP front-view cameras announced on April 11, 2017. As part of the open, innovative and reliable Renesas Autonomy™ Platform for Autonomous Driving Assistance Systems (ADAS) and Autonomous Driving (AD), the R-Car V3H enables Tier 1 suppliers and OEMs to flexibly implement their assisted driving to highly autonomous driving car road.

Advanced R-Car Recognition Technology

The R-Car V3H SoC focuses on architectural optimization of computer vision processing, supporting all ADAS-related functions from conditional autonomous driving to highly autonomous driving. The R-Car V3H utilizes Renesas' heterogeneous computer vision core concept based on the IMP-X5+ image recognition engine and dedicated hardware accelerators to implement advanced perception functions with algorithms including dense optical flow, dense stereo disparity, and object classification. The integrated CNN IP accelerates deep learning with industry-leading low power consumption of only 0.3W, achieving 2x the deep neural network performance of the R-Car V3M.

Scalability and Cost Savings

Based on the proven IP used in the R-Car V3M, the R-Car V3H uses a dual-channel Image Signal Processor (ISP) to convert camera sensor signals for image building and recognition processing. This reuse of IP ensures scalability from NCAP systems using R-Car V3M to Level 3 and Level 4 smart camera systems using R-Car V3H, saving development time while eliminating the need for systems to have every camera ISP, and reduce costs. To further save system cost, the R-Car V3H uses only a single LPDDR4 memory, thus reducing the cost of memory components compared to other front-view camera solutions.

Open Platform Solutions

Tier 1 suppliers and OEMs can choose to develop their own front-view camera solutions or work with leading partners from the Renesas autonomy ecosystem, which is supported by a number of leading front-view camera partners including Hella Aglaia.

Jean-Francois Chouteau, vice president of Renesas Electronics, said: "The specification and design of the R-Car V3H was accomplished through close collaboration with the market leader in front-view cameras to ensure that it meets the needs of leading innovative technologies for autonomous driving systems. As an automotive semiconductor An industry leader, Renesas Electronics is committed to providing open, innovative and reliable solutions for assisted driving and autonomous driving. In addition to achieving higher computer vision performance at a very competitive system cost, our customers use R-Car What I especially liked with the V3H was the ability to retain the freedom to implement a front-view camera with its own unique differentiated technology, while still benefiting from a solution that was scalable from the R-Car V3M to the R-Car V3H.”

supply

Samples of the R-Car V3H SoC will be available from the fourth quarter. Mass production is scheduled to begin in the third quarter of 2019. (Availability dates are subject to change without notice.)

Renesas will showcase the new R-Car V3H at embedded world 2018 in Nuremberg, Germany, from February 27 to March 1, 2018, in Hall 1, Booth 310.