RADAR is an established technology, but interest has been stimulated recently by demands of driver assistance systems and emerging self-driving cars for applications including proximity warning, blind spot detection, adaptive cruise control, and emergency braking. Advanced driver-assistance systems (ADAS) and autonomous driving (AD) systems typically combine several types of sensors, such as cameras, RADARs, and LiDARs. Different types of sensors have their strengths, and effectively complement each other. Cameras and the appropriate machine vision algorithms can “see” lane marking and recognize traffic signals
and signs. LiDAR can offer high (cm-level) resolution and high density of collected data points. RADAR technology is indispensable in ADAS applications because of its robustness to a variety of environmental conditions, like rain, fog, snow, and its ability to directly and precisely measure range and velocity. Basic use cases can rely on RADAR sensors only, enabling cost-effective solutions. RADAR can be used to augment, cross-check or ‘fuse’ situational models derived using advanced computer vision algorithms.
This paper provides an overview of digital signal processing algorithms typically used in frequency-modulated continuous wave (FMCW) RADARs, reviews system tradeoffs, offers a way to approximately estimate computational complexity with a few numerical examples, and finally presents implementation options and solutions available from Synopsys.