What are the opportunities and challenges facing automotive lidar technology?

It is almost a consensus in the industry that LiDAR is one of the important hardware configurations for realizing the advanced intelligent assisted driving (ADAS) function of automobiles. Without LiDAR, there will be no true autonomous driving.

In recent years, Chinese OEMs have used LiDAR technology to revolutionize the automotive industry. As Chinese OEMs take the lead in integrating LiDAR, Yole Intelligence, a subsidiary of Yole, predicts that the global automotive LiDAR market will grow from US$317 million in 2022 to US$4.477 billion by 2028.

The PC (patrol car, a means of transportation used for public security patrols, pedestrian streets, golf courses, tourist attractions, real estate, garden communities, large enterprises, parks, entertainment venues, stadiums, colleges and universities, hospitals, sanatoriums, stations, airports, docks, etc.) and LCV (light commercial vehicle) lidar markets have taken off, with 144 design introductions since 2018, of which 114 are scheduled to be implemented soon after 2023.

Since the automotive lidar market is still relatively young and market share is changing rapidly, the ecosystem faces considerable uncertainty. China is driving the market forward, and the technology, except for wavelength, is likely to change significantly.

Two major application areas have different trends

Automotive LiDAR is mainly used in two fields: PC and LCV and Robotaxi (driverless taxi), each of which has different development trends. Since 2019, Robotaxi's share in this market has been higher than that of the passenger car market. Of course, the LiDARs used are also different. The former is an expensive LiDAR, with each LiDAR costing as much as $75,000 initially; the latter is a LiDAR developed by Waymo and other manufacturers after 2019, with a 90% reduction in cost.

In fact, as early as August 2016, NuTonomy launched its first Robotaxi service in Singapore, followed by Waymo, a subsidiary of Google's parent company, and Cruise, a subsidiary of General Motors, in 2017. The operations of these two companies began in Phoenix, Arizona and San Francisco, California, respectively.

Today, Chinese Robotaxi companies have emerged around the world, with Didi, AutoX and Baidu all joining the race. These companies are launching new services in cities around the world, making the LiDAR market for Robotaxi far ahead of the market for passenger car applications.

“We are at a crossroads in 2022, when the PC and LCV and Robotaxi markets generate almost the same revenues: $169 million for passenger cars and $163 million for Robotaxi,” explains Pierrick Boulay, Senior Analyst for Lighting & ADAS Systems at Yole Intelligence. “However, the two markets have different growth potentials: passenger cars will grow at a CAGR of 69% between 2022 and 2028, while Robotaxi will grow at a CAGR of only 25% during the same period.”

2022-2028 Automotive LiDAR Market

Chinese OEMs gain momentum

It is clear that Chinese OEMs are pushing for LiDAR to be integrated into their cars, with more than 25 different Chinese OEMs currently equipping their cars with LiDAR.

While OEMs in the EU or US have limited the use of LiDAR to F-segment cars (full-size luxury models), Chinese OEMs are now launching D-segment cars (large family cars) equipped with LiDAR. These cars are much cheaper than the F-segment segment, and since they are produced in much larger quantities, the use of LiDAR will also surge.

The automotive lidar ecosystem is quite large, with 50 companies in the spotlight, some of which are already in mass production, while others are still in the R&D stage, developing the next generation of lidar.

Pierrick Boulay believes: "As far as the Robotaxi market is concerned, few companies can generate significant revenue. Hesai Technology controls the entire market with a 67% market share. Their lidar is used in most Robotaxi, such as Cruise, Aurora, Apollo, Didi, Pony.ai and AutoX. In fact, Hesai Technology has replaced Velodyne, the global lidar giant that previously led the market."

In the PC and LCV market, the situation is changing rapidly and is expected to change again in the next few years. Since 2018, Valeo has been the market leader due to its partnerships with Audi, Honda and Mercedes, as the company supplies the majority of the cars released with LiDAR.

In 2022, the market has changed dramatically. Innovasion, which was relatively unknown before, set a new record and delivered more than 56,000 lidars to NIO. Innovasion's installation volume is close to Valeo, with a gap of only a few hundred units.

One of the most notable changes is the addition of Hesai Technology, which ranks third, and RoboSense, which ranks fourth, is also following closely. These Chinese companies only started mass shipments in 2022 and have established partnerships with most Chinese OEMs.

Other players shared the remaining 11% market share; DJI Livox’s market share reached 8% due to its partnership with Xpeng.

Hesai Technology and RoboSense are expected to lead the passenger car LiDAR market, occupying the top two positions in 2023. This is a huge success for the Chinese players, but Valeo is still in good shape as it has $1 billion in orders for its third-generation product, Scala 3.

LiDAR market share for passenger cars and light commercial vehicles, 2021-2022

LiDAR technology "change" is the main theme

In terms of technology, there are four categories to watch, of which only one is expected to remain roughly the same over the next decade: wavelength. In the other three categories, we can see major developments in imaging technology, emitter type, and type of photodetector used.

Wavelength is the only thing that remains stable. NIR (near infrared) wavelengths (905/940nm) are expected to account for 84% of passenger car lidar usage in 2023. Over the next decade, the split between NIR and SWIR (short wave infrared) is expected to remain constant.

In the SWIR field, only two types of components can be used: 1550nm fiber lasers and APDs (avalanche photodiodes). The advantage of the 1550nm wavelength is that it is eye-safe and has a longer detection range, but this also comes at a price. The power consumption of this type of lidar is around 30W, while the power consumption of near-infrared lidar is between 10W and 15W.

In terms of imaging technology, hybrid solid-state LiDAR based on rotating mirrors dominates, with an estimated usage of 68% in 2023, followed by LiDAR based on MEMS (micro-electromechanical systems), accounting for 30%. In the next decade, the share of LiDAR based on rotating mirrors will remain at 56%. The share of MEMS LiDAR is expected to drop to 7%, because strictly speaking, MEMS cannot be considered a pure solid-state LiDAR, because the mechanics are not completely eliminated in the MEMS solution, but the mechanics are miniaturized.

The newly emerged all-solid-state flash (flood-array) laser radar may occupy a 32% market share in the next decade. Its principle is to emit a large area of ​​laser covering the detection area in a short period of time, and then use a highly sensitive receiver to complete the image of the surrounding environment. Flash laser radar is a relatively mature technology among all-solid-state laser radars.

The above changes are related to changes in optical emitters and photodetectors. In fact, multi-junction based VCSEL (vertical cavity surface emitting laser) arrays are improving rapidly, and combined with SiPM (silicon photomultipliers) or SPAD (single photon avalanche diodes), they will be able to provide all-solid-state lidar with no moving parts.

VCSEL is a new type of laser that emits light vertically to the surface, and its structure is different from that of traditional edge-emitting lasers (EEL). This brings many advantages: small divergence angle and circularly symmetrical far and near field distribution greatly improve the coupling efficiency with optical fiber, no complex and costly beam shaping device is required, and the coupling efficiency with multimode optical fiber can exceed 90%.

In addition, many physical properties of VCSEL are superior to traditional solutions, which can achieve higher power output while significantly reducing development costs. In terms of manufacturing process, VCSEL is compatible with light-emitting diodes (LEDs), which helps reduce the cost of large-scale manufacturing.

As the name implies, multi-junction VCSEL is relative to single-junction VCSEL. The single-hole optical output power of a single junction is generally 5-10mW. Taking the five-junction VCSEL array released by Lumentum in 2021 as an example, the single-hole output optical power of its 905nm VCSEL can exceed 2W. Experimental results show that the peak power density of the five-junction VCSEL is greater than 2kW/mm at 105°C. At present, multi-junction VCSEL arrays have become a product that major manufacturers around the world have gradually focused on. In the next 5-10 years, VCSELs are expected to gradually replace edge-emitting lasers and become the mainstream laser components.

Especially in autonomous driving applications, the use of dToF (direct time of flight) VCSEL lidar can produce various types of sensors through different array modes, which is expected to completely replace the existing single-function infrared sensors, ultrasonic sensors, and camera sensors with insufficient privacy and security.

In fact, SiPM and SPAD are much more sensitive than APD. These two components are replacing APD in the near-infrared region, and in the next decade, SiPM and SPAD will account for 60% and 22% of the market respectively.

There are three difficulties in installing LiDAR on vehicles

At present, the large-scale installation of LiDAR in vehicles still faces many difficulties. The first is the high cost; the second is the need for longer detection distance, higher resolution, and higher signal-to-noise ratio; and the third is lightweight and all-solid-state. In terms of core light sources, the use of low-cost VCSEL is the key to achieving this goal.