Deep Learning Enables Fast System Development For ADAS Applications
New software development kit (SDK) is an easy-to start and use development framework for the R-Car V series SoCs
Automakers are increasingly relying on deep learning to enable smart camera applications and automated driving systems for next-generation vehicles. However, most deep learning solutions available today are built on consumer or server applications, which do not operate under the same stringent constraints for functional safety, real-time responsiveness and low power consumption.
To address this while enabling quick and easy software development and validation for smart camera and automated driving applications used in passenger, commercial and off-road vehicles, Renesas Electronics Corporation has introduced the R-Car Software Development Kit (SDK).
Optimised for use with Renesas’ state-of-the-art R-Car V3H and R-Car V3M system-on-chips (SoCs), the new R-Car SDK is built for rule-based automotive computer vision and AI-based functions. Its simulation platform offers AI and conventional hardware accelerators highly accurate simulations in real-time.
A full suite of development PC-based tools is delivered for both Windows and Linux as well as multiple libraries, including support for deep learning, computer vision, video codecs and 3D graphics. Besides Linux, the new SDK also supports multiple ASIL-D-compliant operating systems (e.g., QNX, eMCOS and INTEGRITY) in a single package.
e² studio for R-Car Integrated Software Development Environment
Renesas offers a special version of its e² studio for the R-Car V series regarding high-performance, real-time computer vision applications in ADAS and automated driving. The open-source Eclipse-based development environment includes a full set of debug features and a cutting-edge e² studio GUI that allows easy customisation and integration of third-party tools. It also supports bus monitoring and debug functionalities for image processing and deep learning subsystems.
Easy to Start, Access, Use and Develop
The new R-Car SDK also includes a full set of software samples, popular CNN networks, a workshop and application notes, which is helpful for beginner users of Renesas R-Car platform to access the technology and enable quick development. It is also an ideal platform for benchmarking Renesas products and select the most appropriate SoC for customers’ target applications.
Thanks to the automatic installer, all the software libraries and the development environment can be launched quickly on a development workstation. Applications developed and designed on a PC can be seamlessly ported to embedded development hardware. Finally, Renesas’ R-Car partner ecosystem – the R-Car Consortium –will have access to the R-Car SDK for integration into their R-Car solution roadmap.
Customers can also combine the R-Car V series and SDK with Renesas’ high-performance, low-power RH850 microcontrollers and automotive power management products for further streamlining hardware design that reduce functional safety validation requirements and power consumed by smart cameras and other computer vision-based applications.
“Software development and delivery have been a significant pain point for automotive system developers, involving resource-intensive customized packaging and full installations that typically take several days to complete,” said Naoki Yoshida, Vice President, Automotive Digital Products Marketing Division at Renesas. “To alleviate these headaches when it comes to deep learning for automotive systems, Renesas is reinventing the developer experience, offering this new single package, multi-OS software platform that is easy for customers to access, learn, use, and install, enabling customers to quick start their deep learning development.”
The R-Car SDK is available now from Renesas Electronics Corporation.