Renesas tackles autonomous driving with scalable platform
Just one week after Mentor Graphics introduced a computing platform designed to serve as the brains of a self-driving vehicle, Renesas has launched its alternative concept. The Autonomy platform is designed as an open ecosystem, allowing system manufacturers to choose their preferred degree of individual position in the “make vs buy” field, providing them the option to select the most suitable IP cores and hardware accelerators for their specific needs – all this taking into account functional safety as an indispensable prerequisite of automated driving.
Autonomy is built around Renesas’ SoC and MCU roadmap, and at the opportunity of revealing the platform, Renesas already populates it with a first SoC that fits into this framework – the R-Car V3M. This SoC is focused on smart camera applications, one of the front-end elements in the architecture of automation (sense – decide – control). Equipped with different accelerators, including a pipeline engine (IMP) and a computer vision engine (CVE), the V3M is also capable of handling other type of sensor signals such as radar or lidar.
The SoC also contains a dual ARM Cortex A53 microprocessor for application programming and dual Cortex R7 lockstep cores to run Autosar software and meet functional safety requirements arising from the safety standard ISO 26262. The integrated image signal processor (ISP) enhances the raw image quality from the camera sensor and prepares the image data for downstream processing in computer vision applications. Another feature of the R-Car V3M is its low power consumption, owed to the requirements arising from its typical position next to a car’s windshield and thus exposed to sunlight.
The introduction of the Autonomy platform and the R-Car V3M results from the finding that Advanced Driver Assistance Systems (ADAS) gradually evolve to autopilots, and each step requiring much higher computing power than the previous one. In this evolution, the governmental car safety program NCAP (New Car Assessment Program) is setting the pace, because automotive OEMs implement those driver assistance systems that are required by NCAP to achieve the highest safety rating (5 stars). And while the evolution from driver assistance system to self-driving systems is a gradual one, the next step is going to be a challenge because the automotive industry is currently preparing the transition to new E/E overall architectures, explains Uwe Westmeyer, Senior Principal Engineer in Renesas’ Global ADAS Center. The current situation in terms of engineering: ECUs are combined in domains and get virtualized; client/server topologies are being introduced. Software reuse is becoming an important topic for several reasons: It helps developers to speed their designs and cut time-to-market. And higher degrees of autonomy are requiring higher levels in the ASIL scale of the ISO 26262 standard. “High performance is one thing; high performance plus Functional Safety is a different challenge”, Westmeyer comments. As an answer to this challenge, Renesas has designed the Autonomy platform and will further develop it in the future. “What we are showing is not a prototype but a near-series implementation”, Westmeyer said.