The future of ADAS
The increasing amount of automotive electronic hardware and software requires significant changes in today’s automobile design process to address the convergence of conflicting goals:
· Increased reliability
· Reduced costs
· Shorter development cycles
The trend is shifting from distributed ADAS electronic controller units (ECUs) to a more integrated ADAS domain controller with centralized ECUs. This means that we are currently at what SAE International designates as Level 2 (Partial Driving Automation), where the vehicle can control both steering and accelerating/decelerating but falls short of self-driving because a human sits in the driver’s seat and can take control of the car at any time.
Shifting toward fully autonomous cars—vehicles capable of sensing their environment and operating without human involvement—requires an increase in the electronic architecture of these vehicles.
With the increase in electronic architecture comes an increase in the volume of data. To handle this data, the new integrated domain controllers require higher computing performance, lower power consumption, and smaller packaging.
The adoption of 64-bit processors, neural networks and AI accelerators to handle the high volume of data requires the latest semiconductor features, semiconductor process technologies, and interconnecting technologies to support ADAS capabilities.
The reduction of electronic modules leads to centralized computing architectures, requiring critical automotive building blocks, including processors with vision processing capabilities, neural networks, and sensor fusion. All while addressing the need for quality, safety, and security.
Every aspect of the car is designed to be more connected, requiring subsystem and SoC designers to expand the scope of safety measures beyond the traditional steps taken to ensure physical safety. Applying the latest embedded computer vision and deep learning techniques to automotive SoCs brings greater accuracy, power efficiency, and performance to ADAS systems.
