LiDAR, the autonomy enabler

An article by Pierrick Boulay, Technology & Market Analyst, Solid-state Lighting, and Alexis Debray, PhD, Technology & Market Analyst, MEMS, Sensors & Photonics, both from Yole Développement (Yole), for SEMICONDUCTOR DIGEST | LiDAR technology was envisioned in the 1930s as a way to use light to explore the atmosphere and clouds. It was before the invention of the laser, a key component for LiDAR, which will be invented in 1960. Thanks to the invention of the laser, LiDAR found applications in space exploration and topography in the 1960s. In 1978, Johannes Riegl founded Riegl, a major topographic LiDAR company in Austria. LiDAR then expanded in wind and industrial applications such as factory automation. A major event happened in 2005 for the LiDAR industry with the invention of the 3D real-time LiDAR by David Hall from Velodyne. This was the start of a major movement in LiDAR with new applications envisioned in automotive and logistics.

High expectations in ADAS applications

Advanced driver assistance systems (ADAS) have been developed for more than ten years now, in pursuit of better safety in the automobile world. Combining a set of sensors, mostly radars and cameras together with powerful ECUs, the technology offered promising breakthroughs at the turn of the last decade. Now, with the development of automated driving functions, LiDAR can also be embedded by some OEMs.

The use of LiDAR in ADAS vehicles started three years ago with the release of the Audi A8, integrating the first generation of the Valeo SCALA LiDAR. At this time, Audi was promising the first ever car with autonomous level 3 capabilities but gave up recently due to the absence of regulation and liability issues in case of accident. Nonetheless, Audi is still offering the LiDAR as an option for different models to improve level 2 driver assistance systems.

Apart from Audi, other carmakers showed their interest in developing level 3 functionalities. This resulted in partnerships with LiDAR manufacturers. Audi remained the only OEM offering LiDAR until recently when Mercedes equipped its recent S-Class with the second generation of the Valeo SCALA LiDAR. This is for vehicles already available. With regard to expected vehicles, several official partnerships have been announced: Volvo and Luminar, BMW and Innoviz, Hyundai and Velodyne, and Great Wall Motors and Ibeo. All of these partnerships have several similarities: LiDAR will be implemented in luxury or high-end vehicles, thus generating low volumes and the LiDAR will be mounted on the front of the vehicle, in a central position, and these vehicles will be released in the next two years.

In this context, it is expected that the LiDAR market for ADAS vehicles will grow from US$39 million in 2020 to US$1.7 billion in 2025 at a CAGR of 114%. This growth could be explained by several reasons. The first is the price reduction of these sensors related to the strategy of companies like Velodyne willing to stimulate adoption through lower pricing. Another is the emergence of short-range flash LiDAR, like the one from Continental. Many of these short-range LiDAR could be embedded in vehicles to enable automated lane change in a highway driving scenario.

In terms of applications, various can be envisaged and all have an impact on the position of the LiDAR on the vehicle and on its specifications. OEMs will have to clearly define the use cases they want to achieve. One possible use case could be highway driving. This is one of the simplest as there is no oncoming traffic and all the cars are driving in the same direction. For this use case, a long-range LiDAR will be used, and it will be placed in a central position, almost aligned with the forward ADAS camera and the long-range radar. If OEMs want to implement an automated lane change, the situation becomes more complex. There will be a need to monitor vehicles from the rear as well as on the sides. That would make three more LiDARs in addition to the first one. These additional LiDAR should be short-range and have a large field of view to avoid any dark zones. Another use case would for city driving, which would be even more complex due to other vehicles but also due to other road situations. In this case, a minimum of four to six short-range LiDAR would be necessary to cover the surroundings of the vehicle.

Regarding the technology used in LiDAR for ADAS vehicles, all of the previously mentioned companies are involved in Time-of-Flight (ToF) technology whether it mechanical, MEMS scanning or flash. Other technologies, like FMCW or phase shift, are expected in the longer term and are currently still in a demonstration phase at best. On the component side, the majority of LiDAR manufacturers are using edge emitter and avalanche photodiode (APD) at 905nm due to the large availability of these components. One exception is Luminar which is using fiber laser and APD but at a wavelength of 1550nm. The challenge for them will be to reduce the cost of their LiDAR to be competitive and achieve the target of US$500 they have set.

But not only edge emitters and fiber laser can be used in LiDAR. VCSELs are being increasingly adopted by LiDAR manufacturers, mostly for short-range even if some have achieved long range reach. In this case, two players are differentiating from others. Ibeo has already a design-win with Great Wall Motors in China with vehicles expected to be released in 2022. The other player is Ouster that has recently announced a LiDAR for ADAS vehicles using a combination of VCSELs and SPADs with a start of production (SOP) in 2024.