June 20, 2024


Frode Tyholdt, PhD

Deep dive: 3D ultrasonic sensing in air

Frode Tyholdt, PhD, explores the long lines of technology evolution that today enables advanced 3D ultrasonic object detection in air. Find out why Sonair's strength is to combine its advanced signal processing expertise with cutting-edge piezoelectric MEMS transducers.

Sound waves are commonly associated with their propagation through air. However, ultrasonic waves, those beyond human auditory perception, have traditionally found their utility in imaging applications within liquids or solid tissues such as water, concrete, or body organs. In air, only distance measurements have been available. Now, advancements in sensor technology is shifting the paradigm of air-based ultrasound.

Norway has established itself as a hub for ultrasound innovation, leveraging robust academic-industrial partnerships. The Norwegian University of Science and Technology (NTNU) and the University of Oslo (UiO) have been pivotal in advancing ultrasound applications for medical diagnostics, SONAR, and oil exploration. Companies in these categories include Kongsberg and GE Vingmed Ultrasound. This ecosystem has birthed groundbreaking technologies, now extending to ultrasound in air.

Leading the charge, Norwegian companies like Sonitor, Elliptic Labs, and Sonair are at the forefront. Sonitor specializes in real-time indoor positioning within healthcare settings, while Elliptic Labs uses ultrasound for proximity detection. Sonair is advancing 3D object detection in air, a significant leap forward in ultrasonic technology. All these companies have in common that they originate from the ultrasound community at UiO.

The mechanics of 3D object detection

To be able to do accurate 3D object detection with ultrasound, you need to create an array of transducers that are spaced half a wavelength (λ/2) apart. This allows you to both send directional ultrasound and also determine the direction the ultrasound is coming from (beamforming). This is something that has been developed for a long time in areas such as medical ultrasound, but not for in-air applications. Much of the reason for this is that the wavelength in air is quite short since the speed of sound is significantly lower than in tissue (343 vs 1540 m/s). This means that the distance between the elements in the aforementioned array must be correspondingly short, and this has been difficult to achieve.

You have long been able to buy ultrasonic transducers for air applications at 40 kilohertz (kHz). However, at 40 kHz, half a wavelength (λ/2) is only 4.3 millimeters. This means you need transducers that are smaller than this to do beamforming, and thus 3D object detection in air. This has not been, and still is not, openly available on the market.

The MEMS revolution

For ultrasound in air, access to MEMS (Micro ElectroMechanical System) technology is crucial to be able to make transducers small enough. The smallest transducer available on the market is 5.2 millimeters, so it is not small enough to do beamforming at 40 kHz. With MEMS technology, however, transducers can be made significantly smaller while maintaining a high much sound pressure. Such a transducer chip can easily be made with a size of 2 mm, which corresponds to (λ/2) at around 86 kHz. This allows for beamforming and thus 3D localization up to this frequency.

Sverre Holm, a professor at the University of Oslo, has worked with ultrasound all his life. When he was introduced to this MEMS technology, his comment was, "Wow, now we can finally do beamforming in air."

Norway has a recognized environment at SINTEF MiNaLab that has worked with piezoelectric microsystems (piezoMEMS) and piezoelectric MEMS transducers since 2002. This is the environment Sonair spun out from.

Frode Tyholdt at the SINTEF MiNaLab

Overcoming historical barriers

The history of ultrasound is closely linked to piezoelectric materials. These are materials that expand if you apply an electric field to them, and they can do this very quickly. When you get an ultrasound check-up at the doctor’s, such a material produces ultrasound at a frequency of several MHz.

The most well-known piezoelectric material, PZT, was first made in 1952. After this, the ultrasound-based equipment we know today was developed, such as medical ultrasound and SONAR. These methods use what are called bulk transducers. This means the transducers are made of 100% piezoelectric material. This gives very good properties, but it also means there is a limit to how small they can be made. At the same time, in the 1960s, integrated circuits based on silicon were developed. This is the technology that all of today's electronic chips are based on.

Ten years later, the same methods used to make integrated circuits were used to create small mechanical systems. This technology is called MEMS and it enables chips with mechanics on a micrometer scale to, for example, create sensors. You probably use several MEMS chips every day. You have, for example, quite a few in your phone. It is perhaps not surprising that miniaturized ultrasound transducers can be made with MEMS technology.

So here we are, over forty years later. Why so long? The short answer: it is very difficult to integrate a 1-4 µm thin layer of a piezoelectric material with silicon, so there was no place you could get chips produced.

Piezoelectric MEMS transducers have been researched quite a bit at universities and research institutes. There was a lot of heavy research on material technology and MEMS transducers from 1990 to around 2010. There have been challenges on several different levels. First, stable production processes had to be developed, and quite complex material technology had to be controlled. After this, the processes had to be stabilized so that these PMUT transducers could be produced at a MEMS chip factory.

The pieces are falling into place

In the last ten years, what is needed for technology adoption has fallen into place. The basic technology has been developed over many years, and better material technology has led to high performance and stable production processes. This means that the components produced perform much better than before. This results in new applications using piezoMEMS as the basic technology. If you want to produce in really high volumes, many large chip factories in the world, such as TSMC, Global Foundries, Silex, and ST Microelectronics have developed the capability to produce piezoMEMS. This allows for volumes for the consumer market.

SINTEF MiNaLab is such a place and has long experience with the production of MEMS chips. MiNaLab started piezoMEMS in 2002 and is today a world leader, especially in developing new concepts. SINTEF has over 20 years of experience and well-established processes.

Sonair’s strength lies in its integration of advanced signal processing expertise with cutting-edge piezoelectric MEMS transducers. By harnessing these technologies, Sonair is redefining the possibilities of 3D ultrasonic sensing in air, paving the way for new applications and advancements in robotics, software, and system integration.

Header picture courtesy: Lisbet Jære.

Keep reading