3D printer Lüneburg lens replaces radar and environmental sensors

Autonomous driving
3D printer Lüneburg lens replaces radar and environmental sensors

presenters on this topic

In order to drive a car automatically, approximately 15-20 sensors are currently required. The American company Lunewave sees it differently: four of its radars will be enough for Level 5 autonomous driving. What is behind it?

That’s how cool golf ball-sized radars look from the Lunewave. But they pack a punch: The Lüneburg lens consists of 6,000 chambers. Their production can only be carried out in 3D printing.

(Photo: Kyle Keener/The Vosgard Group)

The radar sensors currently in use differentiate things, but they don’t realize if they’re alive or not. In the case of an automatic evasive maneuver, this can be critical. So Lunewave’s radar system should be able to detect this important difference. “Our algorithms can tell if a person is standing there or something is not alive,” said Hao Xin, Lunewave’s chief technology officer.

This works because the radar signals sent from a person and a non-living thing show subtle differences. For example, a person always moves a little. Radar “sees” they are breathing and even their hearts are beating. These characteristics, which are characteristic of a living organism, can be recognized by a special radar algorithm. However, the radar must have correspondingly high resolution first.

Four radars only for level 5

Level 3 self-driving cars are already on the road in Japan. They mostly use radar. The system usually consists of a long-range radar at the front and four short-range radars. With standard radars, five devices are needed to reliably send a Level 3 vehicle onto the road.

Photo Gallery

Photo gallery with 5 photos

According to Lunewave, you only need four cars to drive to level five fully automatically. Because the spherical antenna of your Lüneburg lens is designed in such a way that the incident rays are focused at a point on the other side. This ensures very good reception quality for all directions in the room. They also work from the receiver to the transmitter: a focused signal is transmitted as a wider wave.

The Lüneburg lens covers a radius of 360 degrees – provided the radar is mounted on the roof of the vehicle. At the moment, Lunewave only conducts tests using the radar on the hood or bumper. It hits 180 degrees horizontally in position, because your vehicle is blocking the remaining radius.

The radar currently perceives an angle of up to 30° on elevation, thus all hills, bridges, signs, etc. that the vehicle encounters on the road. Looking into the future of drones and flying cars that carry passengers, Lunewave operates on a 180-degree vertical radius.

The complex structure of the Luneburg lens

The Lüneburg lens antenna is round. For an antenna, the dielectric constant of the material, permittivity, should be two. The constant indicates the ability to polarize a substance by an electric field. Near the edge of the antenna, it should constantly drop to the value 1.

Complex Street Scene Simulation: Radar Scene Simulator simulates the real world.  This allows the radar sensors to be extensively tested in the laboratory.

Since the antenna is spherical, the structure starts in the center and continues layer by layer towards the edge. In doing so, the permittivity of the material should decrease. To achieve this, the layers become thinner and thinner. The antenna has at least ten layers, with some high-quality products up to 100 layers.

The frequency should be higher, the layers should be thinner. Due to the difficulty of producing this, only very expensive Lüneburg lenses have been available so far. Conventionally manufactured lenses currently have a maximum limit of ten gigahertz. Car radar needs 77GHz.

Using 3D printing for higher radar frequency

In order to achieve the necessary frequency, the thickness of the material layers in Lunewave is 0.1 mm; like paper. “It is impossible to produce by conventional methods. That is why we use additive manufacturing,” says Shen.

Simply printing very thin layers does not result in a Lüneburg lens. Lunewave breaks down the CAD antenna design into thousands of tiny blocks. Each of these blocks gets a different geometry. The entire structure is very complex. Shane is convinced that “only 3D printing can produce the lens.”

It is printed by photopolymerization – more precisely by stereolithography and polymer jetting. The two resin-based processes represent smooth surfaces and fine detailing. Lunewave uses PMMA standard acrylic material as material.

With additive manufacturing, the field of application of technical ceramics is expanding.  For example, the so-called Lüneburg lens is used for communication between self-driving vehicles.

Production – a big secret

Although Lunewave uses a standard printer, it has been significantly modified. “The accuracy and design of our antenna is complex – particularly at high radar frequencies. We had to develop our printing protocol and design the post-processing accordingly. This would not work with a standard 3D printer because the accuracy and complexity is too great. This even exceeds some 3D printers High quality,” Shane describes. The entire manufacturing process from pre-treatment to post-treatment remains a major trade secret.

On the other hand, the electronics of the system are on the shelf. “Right now, we’re using silicon off-the-shelf integrated circuits,” says Shen. All data can already be processed using existing signal processor chips. The company can currently print 100 antennas per day. Even with a single industrial printer designed for antenna needs, it would be 1,000 lenses a day. One of the benefits of 3D printing is that the process is linearly scalable. This means that with ten printers, 10,000 Lüneburg lenses can be produced per day.

That looks promising. But what about the terms of use? Is the radar keeping up with them? Since it is an electronic component, there are many tests proven to check the quality under real world conditions. “The 3D-printed antenna has passed all qualification tests in the automotive sector. Materials, lens, mechanical structure, everything that can be used safely in the car,” says Shane proudly.

Why don’t you want a Tesla Radar?

Radar emits electromagnetic waves that are reflected by objects and received by the radar. These “echoes” provide information about objects. Lighting conditions, poor visibility, dust, rain, day and night, or direct sunlight do not affect the radar measurements.

However, Tesla does not use radar. why? There were incorrect measurements, which led to phantom braking. According to British market research institute IDTechEx, Tesla used the Continental ARS4-B radar – a device from 2014 with outdated technical capabilities. For example, radars at that time could collect information on eight virtual channels. Current Continental radars offer 192 channels – other manufacturers promise more than 200 to 2000 channels. But this technology is also more expensive in return.

(No.: 48256824)

Leave a Comment