The price is only 1/10, lidar killer comes on the scene

In-vehicle sensors , a new revolution has begun.

Only forefoot and head lidar players have developed a lidar with a detection range of 300m and a cost that is half that of mainstream solutions.

Today, 4D imaging radar with more price advantages is coming to the market.

Not only can it identify pedestrians 200 meters away (yellow is the lidar recognition part):

A car moving 300 meters away:

Even a moving car on a bridge 200 meters away may not be recognized by lidar, but 4D imaging radar can accurately identify it:

The most important thing is that the price of this 4D imaging radar is only 1/10 of the lidar, and it is still the kind that can be produced in large quantities.

Altos Technology, a company from Silicon Valley that was only a year old, surfaced. The founder was a Chinese engineer who worked on Apple’s autonomous driving project and Pony.ai.

What kind of 4D imaging radar?

Altos V2 is the second 4D imaging radar product released by Altos Technology.

Based on the TDA4 chip cascade solution, officials said Altos V2’s performance is not only comparable to lidar, but can even be better in some scenarios.

In terms of detection distance, Altos V2 can detect pedestrians 200 meters away, small cars 400 meters away, and large cars 600 meters away, and these are all actual measurement results.

Moreover, Altos V2 can also achieve instantaneous and accurate speed measurement. It can accurately measure speed in a single frame without the need for multi-frame deblurring.

For example, if a pedestrian is stopping in front of the vehicle, Altos V2 can track the pedestrian's movement in real time.

In addition, Altos V2 also features high angular resolution, high point cloud density and low noise.

Under the 16dB angular dynamic range, the azimuth and pitch resolution of Altos V2 are less than 1.5°, and it can accurately identify low and high objects, such as traffic signs, overpasses, etc.

The high point cloud density and low noise of Altos V2 allow the point cloud to be stably concentrated on effective targets, improving the perception effect.

For example, this is an occupancy network built in real time using Altos V2. It can not only identify the driving lane, but also identify the side roads and highway ramps.

And most importantly, Altos V2 is not only a 4D imaging radar that is already in mass production, but its price is only 1/10 of lidar, which has a great cost advantage.

In terms of cost alone, 4D imaging radar can already pose a threat to lidar.

4D Imaging Radar: Lidar Killer?

Here is a brief introduction to 4D imaging radar, that is, 4D imaging millimeter wave radar .

It is called 4D imaging because 4D imaging millimeter wave radar adds the latitude of pitch angle (height) to the three latitudes of traditional 3D millimeter wave radar that can sense distance, speed and horizontal angle.

Therefore, 4D imaging radar can achieve higher resolution than traditional 3D millimeter-wave radar, output a denser number of scatter point clusters, and then form a point cloud image, providing a similar perception effect to lidar.

The point cloud image generated in real time by Altos V1, the previous generation product of Altos Technology, is compared with the effect generated by Hesai AT128 lidar as shown in the figure.

The difference between 4D imaging radar and lidar is that although both use echo imaging, lidar measures and calculates by emitting lasers and receiving signal echoes; 4D imaging radar emits electromagnetic wave beams and uses the Doppler effect. , detect the frequency difference of the reflected signal, and then perform measurement and calculation.

It is precisely because of this difference that 4D imaging radar can work stably in bad weather, while one of the "flaws" of lidar is that its performance will be weakened in bad weather environments such as rain, snow, haze, sand and dust.

Coupled with the cost advantage, it seems that 4D imaging radar may really be an alternative to lidar?

However, even if 4D imaging radar can replace it in terms of performance, there are still some challenges in actually using it on actual vehicles.

For example, 4D imaging radar has higher requirements on algorithms than traditional 3D millimeter wave radar. Issues such as how to deal with the larger amount of data brought by the increase in point cloud density and how to identify and classify targets more accurately are all required. solve.

In addition, the 4D imaging radar industry as a whole is still on the eve of large-scale mass production, and the overall 4D imaging radar on-board situation can also prove this from the side.

Currently, mass-produced models using 4D imaging radar include SAIC Feifan R7, Changan Deep Blue SL03, Lotus ELETRE and other models, which are not many compared with models using lidar.

This field as a whole is also in a stage of liveliness, chaos, and melee among heroes.

Players laying out this track include traditional Tier 1 players such as Bosch , Continental, Aptiv, ZF, and Huayu Automobile, as well as technology companies such as Mobileye and Huawei , as well as Arbe, Oculii, Senstech, Emerging companies such as Freetech, Muniu Technology, Chengtai Technology, Zongmu Technology, and Geometry Partners.

Whether 4D millimeter wave radar can really become a replacement for lidar is still too early to draw a conclusion, but the possibility still exists.

Who is Aotu Technology?

Finally, let’s briefly introduce the company that created the Altos series of products, Aotu Technology.

The company was established in January 2023. The three founders have extensive experience in autonomous driving, radio frequency design, signal processing, computing optimization and radar perception.

Co-founder and CEO Niu Li graduated from the Automotive Department of Tsinghua University and later received a PhD in Automotive Science from Clemson University.

He once worked for Hitachi and ZF as an automotive electronics /EMC engineer; he was also an early member of Apple’s autonomous car department and an early member of Pony.ai’s Silicon Valley hardware team. He was independently responsible for sensor hardware and led the completion of high-performance 4D imaging millimeter wave radar. Development and testing, vehicle electronic architecture design and other work.

Co-founder and CTO Li Xiaobo holds a PhD from the Institute of Electronics, Chinese Academy of Sciences.

He once served as the head of Huawei HiSilicon radar algorithm and the head of high-performance radar algorithm of Pony.ai.

Michael Wu, co-founder and chief scientist, graduated from Rutgers University.

He once served as the head of system performance for Mozilla (Firefox), an optimization expert for Pony.ai's Silicon Valley plan, and has experience in the autonomous driving departments of Apple and Lyft.

One year after its establishment, Altos V2 launched by Altoo Technology is already its second mass-produced 4D imaging radar. This time it also released the Altos RF series of 4D imaging radar modules without a processor and at a lower price.

The official website introduction also clearly stated their goal: they believe that extremely high-performance and low-cost radar will change the rules of the game in the automotive industry in the next few years.

As mentioned above, in high-end smart driving sensor solutions, it has not yet been determined whether 4D imaging radar can really replace lidar.

However, some players have begun to take sides, and Tesla is still the number one catfish in the smart car industry . On the new Model X launched last year, 4D imaging radar has been installed on the vehicle, making the ultimate purely visual solution for autonomous driving that Musk calls less purely visual.

The probability of 4D imaging radar subverting the smart car sensor landscape is not zero. At least now, in terms of performance and price, 4D imaging radar has enough advantages compared to traditional millimeter wave radar and lidar.

There are still many possibilities in this year’s smart driving industry.