Sensors have always been critical components in the roadmap to autonomy on the roads. This article talks about the different levels in automation, the sensors and then focuses on 4D imaging radar and its recent developments.
Image Credit: metamorworks/Shutterstock.com
According to the standards developed by the Society of Automotive Engineers (SAE), the road towards autonomous vehicles is comprised of five different levels. These standards are generated based on the level of involvement the driver has in the driving process.
Level 0 means that the vehicle does not have any automation involved that aids the driving process. At Level 1, the human driver is still in charge of the driving activities, but the automation system aids in the steering or the acceleration process.
Level 2 is the partial automation stage where the vehicle provides steering and acceleration assist, but the safety-critical functions are still under the driver's control.
At level 3, the automation system in the vehicle starts being aware by understanding the environment and performing dynamic driving tasks such as braking on its own. The human driver needs to be at a capacity to intervene if an unexpected situation occurs.
At level 4, the vehicle is supposed to complete its entire journey without needing the driver to intervene in its operations, even when there is an emergency. The driver will have the option to take over the control of the vehicle if he deems that the autonomous vehicle is not capable of driving in a specific condition.
It's the final level of automation, level 5 automation, that all the automotive companies in the world are ultimately trying to achieve.
At this level, the human driver is classified as a passenger. This level of automation ensures that even in dire circumstances, the vehicle needs to be capable enough to manoeuvre itself from danger without the aid of a human driver.
Sensors for Level 5 Automation
With the roadmap planned for achieving level 5 autonomy in the future, it is of prime importance to define the sensors that will aid the journey. For an autonomous vehicle to be aware, the sensors need to be robust enough to detect and collect information so that accurate decisions can be made.
The three major sensors used in an autonomous vehicle are image, LiDAR, and radar. An image sensor on an autonomous vehicle usually has a 360-degree view which provides information about the traffic conditions on the road.
These cameras can detect objects such as other cars, pedestrians, traffic lights, and traffic signs. The information is then sent to the computing unit, which then analyses and takes the best decision for maneuvering the vehicle.
LiDAR sensors, on the other hand, use the reflections of the lasers to create a three-dimensional image consisting of objects in their vicinity through mapping. LiDAR also provides the distances of the objects from the sensor and can be focused on a narrow or large field of view.
Radar is another prominent sensor that has been remodeled from its military usage to fit in the autonomous driving perception tasks. Radar sensors send out radio waves, and based on the reflections received the sensor computes the distance of the object from itself.
They can be configured to act as short-ranged or long-ranged depending on their purpose and the type of object that they are intended to detect.
Compared to LiDAR technologies, radar can identify objects even in poor visibility conditions.
4D Radar's Environment Response
Due to the host of advantages radars offers compared to their other sensor counterparts, automotive OEM's all around the world have since started modifying radars, making them robust for development purposes. The development of a 4D radar is one such step in this direction.
The 4D radar maps the environment by using echolocation and time-of-flight measurement. The radar is called 4D because besides mapping distance, relative speed, and azimuth, it also uses time for understanding and classifying a stationary object in its vicinity based on its height above the road level.
4D radar consists of a multiple-input multiple-output (MIMO) 48-antenna array for mapping its surroundings. As an output, point cloud data is obtained, which, when combined with a wide azimuth-elevation field of view (FOV), provides tracking efficiency and accuracy in case of extreme situations such as a traffic jam under a bridge.
The Market for 4D Imaging Radar Technologies
There are a host of different companies that are developing the 4D imaging radar software. The key focus of all the companies has been to expand the functionality of radars to replace other sensors.
Vayyar is one such company introducing a "radar-on-chip" technology that consists of AEC-Q100-qualified 48 transceivers with an internal, digital signal processor and a microcontroller unit used for real-time signal processing. The company claims that its sensor can execute complex imaging algorithms without the need for an external CPU.
Nxp and Continental have also partnered together to sample a new radar sensor chipset. Each new sensor chipset is capable of classifying radar echoes based on the type of object on the road. The radar will also have the capability of calculating the elevation, range, speed, and azimuth so that smaller objects are not undetected.
The Future of Radar
With companies like Vayyar & Continental developing versatile radar systems that can eliminate the need for LiDARs and image sensors, the market for 3D imaging radar looks bright. Although currently a niche market, 4D imaging radar modules are expected to dominate the automotive industry in the next few years, with experts predicting that by 2025 and 2030, the market will reach $8 billion and $12 billion, respectively.
Continue reading: The Future of Ocean Health Diagnostics with Argo's Biogeochemical Sensors.
References and Further Reading
Khvoynitskaya, S. (2020) 3 types of autonomous vehicle sensors in self‑driving cars. [online] www.itransition.com. Available at: https://www.itransition.com/blog/autonomous-vehicle-sensors
M, Slovick. (2021) 4D Imaging Radar Looks to Advance ADAS and Level 5 Automation [online] Available at: https://www.electronicdesign.com/markets/automotive/article/21151406/electronic-design-4d-imaging-radar-looks-to-advance-adas-and-level-5-automation
www.businesswire.com. (2021) 4D Imaging Radar in Autonomous Vehicles 2021: ADAS and AV Radar Module Market Size Anticipated to Reach $8 Billion by 2025 and $12 Billion by 2030, at a CAGR of 10.5% - ResearchAndMarkets.com. [online] Available at: https://www.businesswire.com/news/home/20211007005490/en/4D-Imaging-Radar-in-Autonomous-Vehicles-2021-ADAS-and-AV-Radar-Module-Market-Size-Anticipated-to-Reach-8-Billion-by-2025-and-12-Billion-by-2030-at-a-CAGR-of-10.5---ResearchAndMarkets.com