Ole Marius Rindal, co-founder of Sonair answers questions about ADAR, explaining how it differs from LiDAR and what it means for automated packaging and palletizing operations.
What is ADAR?
ADAR (acoustic detection and ranging) refers to a new category of sensor that enables 3D ultrasound imaging in the air. ADAR can be considered alongside established technologies like radar and sonar. The key difference is that ADAR works acoustically in the air, whereas sonar works under water and radar uses radio frequencies.
Doesn’t ultrasound in air sensors already exist?
Yes, but the most common version — the parking sensors on your car — are one dimensional. They sense in one direction and provide distance information only. ADAR-based sensors provide the x, y and z coordinates necessary to create a 3D view. This allows ADAR-based sensors to be used on robots for safe operation in shared spaces with humans.
What advances make this new sensor possible?
Beamforming — the principle that underlies ADAR and well-established ultrasound technologies such as 3D medical ultrasound imaging and SoNAR (sound navigation and ranging) — allows sensors to transmit and receive sound from different directions. But beamforming in air requires transducers that are small enough to be spaced half a wavelength (?/2) apart.
A team of Nordic researchers built miniaturized transducers to enable beamforming in air for the first time. The researchers also found that they could apply modified versions of the algorithms used in medical ultrasound imaging on the new sensor, confirming that the new technology could readily perform depth sensing and obstacle detection.
A 2D safety certified LiDAR typically costs about $4,000, whereas the new 3D ultrasound sensor will cost around $1,000 when it’s released later this year. With AMRs typically carrying two 2D safety LiDAR, removing these expensive sensors produces an immediate $8,000 saving.
What does this mean for packaging and palletizing applications that commonly use mobile robots?
The new sensor provides robots with 3D depth sensing and obstacle detection capabilities with a range of up to five meters (16.4 ft). Moreover, adding four of the new, low-cost sensors to an AMR enables 360-degree sensing.
For humans, knowing that the robots around them possess 360-degree obstacle detection provides peace of mind. And, since robots fitted with ADAR technology are built to avoid objects in their environment, this means less unexpected downtime and improved process flow — whether that’s in a warehouse or a manufacturing facility.
Consider the autonomous mobile robot (AMR) that delivers items to the busy packaging or palletizing area, or the collaborative robot that helps with packaging applications: both operate with humans moving around and can benefit from adopting the new sensor technology to ensure safe operation.
The technology could also be applied to entire robot cells, ensuring safety not just in a 5 m range around a robot, but across an entire automated packaging or palletizing cell.
What types of robots can this technology be used on?
The new sensor has been tested extensively with AMRs with excellent results. But it is by no means limited to AMRs. Any robot that shares space with humans can benefit from 360-degree obstacle detection.
Don’t the sensor packages on AMRs already provide obstacle detection capabilities?
Sensor packages, which can account for 30% or more of the total hardware cost of an AMR, incorporate various combinations of sensing and camera technologies — such as 2D safety certified LiDAR — to help robots navigate their environment while safely detecting obstacles.
However, LiDAR has a limited field of view and can only sense a small, two-dimensional slice of the environment around a robot. This means that LiDAR cannot detect objects outside the 2D plane.
Moreover, 2D safety LiDAR are typically mounted on AMRs at a height of around 20 cm (7.9 in). This means that 2D safety LiDAR won’t detect objects that are 10 cm above the floor. Similarly, 2D safety LiDAR won’t detect objects hanging from the roof or items sticking out from walls or shelves if they are outside the sensor’s limited field of view.
Lighting conditions can also create additional challenges for LiDAR, as do transparent and shiny surfaces, which LiDAR cannot detect.
The new ADAR-based sensor provides enhanced obstacle detection versus 2D safety LiDAR.
How does the cost of ADAR compare to alternatives like 2D safety LiDAR?
The new 3D ultrasonic sensor is considerably cheaper, and it can be mass produced. A 2D safety certified LiDAR typically costs about $4,000, whereas the new 3D ultrasound sensor will cost around $1,000 when it’s released later this year. With AMRs typically carrying two 2D safety LiDAR, removing these expensive sensors produces an immediate $8,000 saving. Deploying four ADAR-based sensors instead, at an approximate cost of $4,000, provides the AMR with a full 360-degree protection from obstacles at 50% lower cost compared to 2D LiDAR-based alternatives.
Don’t 2D safety LiDAR sensors support other functions besides obstacle detection?
Yes. 2D safety LiDAR are also used to support autonomous navigation capabilities, for example. To retain those capabilities, robot designers could use the new sensor with much cheaper, non-safety rated LiDAR, which costs around $500, while the new sensor performs the improved 3D safe obstacle detection function.
Depending on the configuration, the new ADAR-based sensor technology reduces the cost of the typical safety sensor package between 50% and 80% while also boosting safety.
What about the fact that 2D safety LiDAR is considered an optimal sensor for AMRs because it produces millions of data points when performing obstacle detection functions, whereas ADAR produces a tiny fraction of that number?
The idea that creating more data points necessarily leads to improved safety is a misconception backed up by participants in our ongoing early access program. Participants in this program include:
- Wheel.me: a supplier of AMRs to the automotive industry.
- Solwr: a manufacturer of AMRs for picking in retail environments and a giant pallet sorter robot.
ADAR-based sensors in combination with advanced signal processing algorithms provide actionable data points to detect obstacles surrounding a robot without requiring millions of data points to function. This means less on-board processing requirements for the robot.
Since LiDAR is so well-established in the mobile robot arena, what do you tell automation professionals who will likely by cautious of this new technology?
Before trying the technology, some people have concerns about using ultrasound in air, because it hasn’t been done before. After using the new sensor, those concerns dissolve quickly.
For example, some people expressed concern that ultrasound would be too sensitive, picking up every object in its vicinity. However, advanced signal processing algorithms filter out unnecessary information, ensuring that the new sensor detects obstacles effectively.
Other people had tried older ultrasound technologies and been unimpressed with the results. However, the combination of miniaturized transducers and process algorithms make the new sensor a different proposition on both the hardware and software side.
I’m not suggesting that this new 3D ultrasound sensor will replace all the other sensors used on robots today. There is far too much variety in automation applications to make such a claim. The future is one in which this new sensor will be combined with others to ensure that humans and robots can safely coexist.
