Software has been developed specifically to use the hardware accelerators in Renesas Electronics’ R-Car V3H SoC for advanced driver assistance systems (ADAS) in automotive design.
To accelerate the development of ADAS, the Perception Quick Start software, based on the R-Car V3H SoC delivers reference software for camera obstacle detection (COD), lidar obstacle detection (LOD), and road feature detection (RFD), deemed as three key recognition areas for sensor-based Level 2+ autonomous vehicle systems.
The COD reference software uses convolutional neural network (CNN) IP, a computer vision engine (CV-E), and image rendering (IMR) technology to detect 2D objects such as cars, trucks, buses, and pedestrians. It achieves approximately 30 frames per second.
The LOD software uses CNN-IP and CV-E to detect 3D objects, including cars and trucks. The LOD achieves approximately 15 frames per second with 3D bounding boxes at 50m.
The RFD reference software uses CNN-IP, CV-E, IMR, and a versatile pipeline engine (IMP) to identify drivable free space, lanes (crossable and uncrossable), road boundaries, and distances to lanes and nearest objects to support NCAP 2020. The RFD achieves approximately 30 frames per second.
The R-Car V3H SoCs deliver a combination of high computer vision performance and artificial intelligence (AI) processing at low power levels, for automotive front cameras in Level 2+ autonomous vehicles. To advance recognition technology, Renesas designed the SoCs with dedicated hardware accelerators for key algorithms including convolutional neural networks, dense optical flow, stereo disparity, and object classification. The Perception software provides an end-to-end pipeline reference for developers working with these complex accelerators which are both cost-effective and power-efficient, thereby allowing customers to advance an application design even if they have limited experience at using the accelerators. The reference software covers input from sensor or recorded data, all stages of processing and display output on a screen.
“Specialised hardware accelerators play an essential role in achieving the computer vision performance and accuracy required in embedded ADAS and autonomy applications while still meeting stringent in-vehicle power consumption limits,” said Tim Grai, director or automotive advanced systems innovation department, Renesas. “However, the complexity of these accelerators can present a steep learning curve. With the Perception Quick Start software, we are able to offer a set of application software along with the underlying primitives to simplify the use of these complex accelerators needed to achieve embedded ADAS.”
Renesas will demonstrate the Perception software at TU-Automotive Detroit (Booth C190, 5-6 June, Novi, Michigan, USA).
