Introduction
In ADAS or AD, the use of multi-sensor fusion can help solve detection, tracking and target identification problems, improve system redundancy and reliability, enhance data credibility and scene coverage. However, in the face of automotive applications, integrating different sensors together still faces significant difficulties due to limitations in exterior design, system structure and space. In this regard, market research company Yole’s market and technical senior analyst, Belgian Tier 2 supplier XenomatiX and Marelli Automotive Lighting company have launched a fierce discussion on the issues of lidar and its integration from multiple angles.
Fusion and integration have already begun
“The use of automotive lidar is still in its early stages, and there are not many OEMs installing lidar on their cars,” said Yole. However, more than 15 partnerships have been established between lidar manufacturers and OEMs. “Yole expects that transaction volumes will increase significantly from 2024-2025.”
Yole predicts in its upcoming “Automotive and Industrial Applications Lidar Report” that the automotive lidar market is expected to grow from $85 million in 2021 to $2.3 billion in 2026, accounting for nearly half of the lidar market in 2026. Previously, the application of automotive lidar accounted for only a single-digit percentage.
“WeChat backstage reply ‘0011’ to obtain the English original text of Yole Automotive and Industrial Applications Lidar Report.”
XenomatiX and Marelli are not suppliers in the same subdivision field, one is “perception” and the other is “lighting”. “XenomatiX is a company that provides true solid-state lidar solutions for ADAS, AD and road digitization,” said the article, “and its designed and built products and software can achieve precise real-time 4D-6D digitization, helping to understand the surrounding environment of vehicles and make automotive applications safer and more comfortable.” “Marelli is one of the world’s leading automotive component suppliers,” the article added, “and its automotive lighting division is a global leader in the development and manufacture of front lamps, tail lamps and lighting electronics. Its product portfolio includes cutting-edge products such as lasers, LED matrices, digital headlights and OLED tail lamps.”
Recently, Marelli reached an agreement with XenomatiX to strengthen and expand its position in the sensor field such as lidar, making sensors the company’s fourth pillar. Specifically, this means “integrating lidar with automotive lighting to make OEM design more convenient.”Frederic Chave believes: “After a period of intensive technological exploration, a fusion and integration era has begun. People are becoming increasingly clear about which solutions will prevail. Multi-beam solid-state lidar is definitely one of them.” He pointed out that the automotive industry’s push for lidar solutions is growing day by day. For safety and reputation considerations, more and more OEMs will adopt a comprehensive AD strategy in specific operating fields, and of course, they will also use lidar to provide autonomy.
The current problem is commercialization, turning surviving solutions at the experimental stage into mature ones. At present, the industry is most focused on providing robust sensing solutions for ADAS L2+, until full autonomy is achieved. China and the United States have the highest acceptance in this market. Regulations elsewhere are also evolving, such as the upcoming launch of the L3 Automated Lane Keeping System (ALKS) by the UK Department of Transport, although the vehicle must still travel at low speeds; by 2022, the German federal government will also allow L4 operation on some restricted public roads.
Market demand is still increasing
Filip Geuens believes that the new crown epidemic has obviously slowed down the development of the entire automotive industry. Most AD projects have been postponed, and budgets have been reduced. More importantly, everyone is struggling with issues related to supply and production restarts, and is still struggling now. As a side effect of the health crisis, the electronic product supply crisis has also had a serious impact.
In addition, due to the new crown epidemic, for health reasons, people are more inclined to travel privately. As more people work from home, less traffic congestion will also lead to a shift from public transportation to more private driving. He said: “The new crown epidemic has affected our business, but the reduction in customer interactions has led to more focus on product development, which is beneficial for long-term development.”
During this period, XenomatiX’s lidar has been optimized continuously, and its volume has been reduced by 3 times. In addition, it only consumes half the energy compared to before, and the number of laser beams has also increased by 3 times. In the industrialization process of solid-state lidar, the establishment of the supply chain is most important to make the product manufacturable and reasonably priced. This is also why they are collaborating with Marelli to explore “heterogeneous” integration.
Challenges of automotive applications
Lidar integration is an important aspect of automotive applications, and there are many challenges. Filip Geuens said that lidar is expected to solve many extreme situations, especially when integrated with radar and cameras. However, this seems to be an unfair competition. From day one, people hoped that lidar would be as cheap as cameras and radar technology 15 years ago. At the same time, people expect it to see the surrounding areas, far away, and all directions with human precision. In addition, lidar is expected to handle all weather and road conditions that interfere with cameras and radar.Therefore, the biggest challenge is to determine what are the fair and reasonable application scenarios for LiDAR, to ensure its proper position in vehicles.
The new LiDAR technology helps ECU make faster decisions and overcome the difficulties of increasing speed or intermingling, such as lane changes or mergers. Before implementing more forward-looking solutions such as cooperative adaptive cruise control (CACC), more accurate information-sharing sensors are needed.
In fact, installing more sensors, including LiDAR, unnecessary protrusions, and the space restrictions of processing body parts, present significant integration challenges in terms of styling. What are the key technologies needed to overcome these challenges? Frederic Chave stated that Marelli, as the initiator of the Smart Corner® solution, can integrate sensors into lighting elements and clusters, using existing production knowledge and occupied area from strategic external components to provide unobstructed vision for sensing components.
Handling adverse conditions is a key challenge for LiDAR. And imaging LiDAR (for example, using CMOS detectors for inherent sensor fusion of LiDAR) is a good way to achieve this goal. Imaging LiDAR can measure distances, as LiDAR should do, and it can also take pictures to check for contamination from mud, scattering light (fog), and deception (other light sources). Marelli’s solution has strong capabilities in dirt detection and direct, efficient cleaning.
Where is the integration a problem?
The grille is currently the first hidden position chosen by OEMs to place LiDAR, but where else can it be integrated, and based on what?
Frederic Chave explained, “Two things are crucial: clear vision and aesthetics. LiDAR should not have unexpected blind spots, and the device should not attract attention.” Currently, LiDAR can be placed in both grilles and external front and rear lighting (Smart Corner). These options can best respond to styling requirements, as well as cost, space, and weight requirements.
– Other locations under consideration for monitoring all directions include various positions on the vehicle.
- As for the laser radar itself, there are many options as long as they meet requirements for clear visibility and aesthetics. For example, the laser radar can be placed on the top or bottom of the B-pillar, behind the rear window, in the side mirror, in the sensor strip under the roof, or in the turn indicator. The size of the module is the basis for ease of integration, and by placing a set of modules around the car, the coverage range of the entire FOV (field of view) can be segmented, similar to the mechanical laser radar used on some car roofs.
- As automotive front and rear lighting become more compact, they are increasingly being used for communication. Are lighting components more suitable for integrating new features such as sensors or small displays?
- Filip Geuens answers: “Although we see that the space for adjustment functions in front and rear lights with higher LED efficiency is becoming less, narrower and wider light strips are more favorable to laser radar. Laser radar does not require a large front window for observation. The light strip can be segmented, and certain areas can be allocated to laser radar. In addition, the heating components can be better optimized.”
- Especially for electric vehicles, they are better able to avoid competition for space with front headlights in the engine compartment. They no longer require air intakes, so the front grille will undoubtedly have more space. As for taillights, historically they occupy less space, and small laser radar modules are easier to integrate with them.
- Grille lighting and animation or extended lighting strips are also growing trends. With the help of an innovative materials supplier, all these surfaces provide more opportunities for integrating smart sensors. Overall, lighting will have a dual function: not only emitting light and signals but also capturing light and signals. There will also be some new possibilities and good business opportunities in this area.
- How to achieve mass application of laser radar?
- Frederic Chave believes that OEMs still need to acquire more professional knowledge about laser radar. It is still too early for laser radar business, although consensus has been reached. “We believe that the business volume will increase. This will happen in the second half of this decade. Initially, the laser radar will enhance ADAS functions. Next, it will play a key role in AD. It took more than ten years for radar to become a commodity, so we expect the use of laser radar will not be less.”
- Filip Geuens said that high-resolution depth sensing is essential in our traffic. Unfortunately, humans make mistakes, and interest in driving also decreases (on a crowded planet, people have a strong desire to travel). Therefore, it is in the interest of the automotive industry to establish 3D sensing expertise that meets their needs as soon as possible.Undoubtedly, LiDAR has strengthened safety and provided necessary redundancy for autonomous vehicles: there will be no compromise on safety. LiDAR technology can provide information in unique situations, which will always help make decisions faster and more robustly in situations where speed becomes faster and complexity becomes higher.
In addition, LiDAR is also applied to other fields, such as trucks and vertical takeoff and landing (VTOL) aircraft. There are many similarities between these applications: they require outdoor sensing under different lighting conditions within a similar range. The cost, weight and size requirements for LiDAR used in drones and cars are very similar. Although autonomous drones are less hindered by pedestrians, they still need precise geometric information to find suitable landing points in congested urban environments and avoid collisions with various infrastructures.
Finally, people’s interest in autonomous solutions in the private sector is growing, some of which have already reached a large operational scale, such as mining, agriculture and yard or port freight transport. They all require LiDAR to operate safely under all weather conditions. After the first integration phase, autonomous truck transportation has now been developed in the United States and Asia, of course, requiring LiDAR for safe operation. For such applications, higher requirements, longer distances, stronger shock and vibration resistance and longer natural life are necessary to reduce any time-consuming maintenance operations.
What technology can laugh last?
Just like the camera, in the beginning, the hardware was the main difference, but after a period of time, the differentiation of software and applications became crucial.
Respondents believe that LiDAR in the mobile travel industry requires deep sensing, and the output of many sensors on the car will realize more intelligent fusion: doing more with the same number or fewer sensors. Those that use non-scanning or moving components can detect the entire scene in “one instant” without being limited by distance or power; at the same time, they use flexible modular designs to provide different perspectives, ranges, resolutions, and frame rates; based on semiconductor laser sources and detectors, they can be batch produced and put on the market in a short time, only such technology has vitality.
This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.