Industry News Details

The Role Of 4D Imaging Radar In Reaching Level Two-Plus Autonomy Posted on : Apr 08 - 2021

Radar is so important in our lives. The ability to bounce radio waves back and forth off objects in the environment has become the backbone of many aspects such as air traffic control and navigation, numerous military applications, meteorology, astronomy and law enforcement.

It is also well known that radar has been a crucial element in advancing the autonomous vehicle industry in the long term and advanced driver assistance systems (ADAS) in the short term. The increasing adoption of ADAS capabilities has been a boon to the radar industry. It is a boon that will continue not only because the market will grow but also due to the fast pace of innovation taking place.

If you remember the original videos of bulky vehicles, pioneering autonomous automakers used combinations of cameras, radar and lidar on their vehicles in an attempt to solve the "basic" problem because each "sensor" has a strong suit. Cameras can detect objects in high resolution but are adversely affected by sunlight (blinding) and darkness and can have contrast and depth limitations. Radar can detect well in all lighting and weather conditions but does not provide good resolution. Additionally, contemporary radar solutions do not solve the basic problem that prevents autopilot accidents—false alarms and the inability to detect and distinguish between stationary and moving objects. Lidar does great when it comes to the detailed and accurate detection of objects in space, but it cannot perform well in bad weather.

4D imaging radar, with a large RF channel array, is a new sensor that can provide a solution to all of the problems just mentioned. The automotive industry is well aware of what 4D imaging radar is going to bring to the table in regard to each of the levels of autonomy (based on the well-publicized L1-L5 scale). The industry has been awaiting improvement in radar technology and understands scalability needs to be a dominant factor, which would go a long way to ensure sensors are "future-proof."

However, there are some additional steps the industry needs to take in order to implement 4D imaging radar. It needs to clearly define and set the bar for the minimum needed performance from all sensors that may end up in the "sensor suite," including resolution, false detection rate, range, latency and sensitivity. Then, the ability to integrate advanced clustering, object bounding, advanced free space mapping algorithms and tracking algorithms—all based on imaging radar—will need to take place. View More