LiDAR – Light Detection And Ranging – is a super-hot technology, required by advanced driver assistance systems (ADAS) and robotic cars. That’s creating a demand – but there are only a few products now available, from Valeo and Velodyne among very few others, with the right specifications and robustness to be implemented in cars. The requested specifications are a moving target and camera module, radar system and car manufacturers are trying to avoid using today’s complex and immature systems. Below we discuss Yole Développement’s opinion on what will happen, which is based on our new report LiDARs for Automotive and Industrial Applications 2018.
First imagined in the 1930s by E.H. Synge, LiDAR is a technology that measures 3D space. Today’s systems emit infrared laser pulses, which bounce back from nearby objects. The time taken between emission and return can be used to measure the distance between the source and the object, a technique known as time-of-flight analysis. Such technology offers machines the type of 3D perception that brings them closer to human beings, and instills in them the ability to function autonomously within our world. That has made LiDAR one of the hottest topics in the fast-moving transportation industry. Almost 100 companies have been created in the past few years to fulfill the LiDAR market’s requirements, especially for two kinds of automotive markets currently evolving in parallel.
Yes, we are talking about Autonomous Vehicles (AV), which are divided into ADAS cars or robotic vehicles. For several months now the automotive market has been steering towards two distinct paths, and it’s important to make a clear distinction between ADAS cars and robotic vehicles. They are not the same, and they shouldn’t be confused.
On one hand, ADAS cars are traditional cars that implement increasingly advanced ADAS. The technique ranges from level 0, which means no ADAS, to level 5, meaning complete autonomous driving (AD). Across the whole range, automotive ADAS has to focus on the reliability, integration and cost challenges of serving a market with sales of millions of units. Cost-effective LiDARs are expected to be mandatory to go beyond Level 3, which is pushing companies to pursue their development.
On the other hand, robotic vehicles are part of a more “industrial” approach, serving “Transportation as a Service” (TaaS). This puts $150k of sensors and computing power on top of cars to make them fully autonomous, using the best available technologies, no matters how bulky, expensive and complex. LiDARs that serve the robotic vehicle market will be mainly driven by performance and availability, and will serve a market of several tens of thousands units by 2022.
The market sizes are in fact totally different between ADAS and robotic vehicles. To generalize broadly, LiDAR will win in the early robotic vehicle market, because it has been ready for many years in the industrial market for mapping/topography and robotic applications. There is no need to redesign these existing products from scratch, but massive diffusion into the ADAS market will need that big technological change.
Lots of hype but only a few know where we’re heading
LiDAR products are available at different physical scales, from a few grams to several tons. In the future, LiDAR will also be produced in mass volumes for automotive markets – but it isn’t that simple. By looking at the different technologies on the market, we can clearly see that the LiDAR market is immature – a ‘work in progress’. Very recent investments of more than $800M over the past three years indicate the ongoing dynamics. The investments are good for technology innovation and production capacity building. The flood of money allows start-ups and manufacturers to prototype and launch pre-series sensors for car and robot-taxi makers who are testing all types of LiDAR internally. It will take some time to gather feedback and determine what the real requirements are, and that’s why the maturation of LiDAR will take a long time. But time is only one piece of the puzzle; diversity in the technology is also the other tricky aspect to understand this market. From big mechanical rotating LiDAR, MEMS micro-mirrors, to optical phased arrays, or flash LiDAR, the landscape of technologies has never been so diverse. This is a complex situation where time-to-market uncertainty and technology diversity prevents any clear-cut vision of which one will win. With average selling prices (ASPs) ranging from several thousand dollars to $65,000, current LiDAR system are still expensive, which cover a broad range of applications that are still niche markets for now. With the emergence of two massive markets to address, the business opportunities are huge, and will drive costs down, and more.
The winners in LiDAR will surely be the ones that better understand market needs. A rather similar situation happened in the past with far-infrared microbolometers. By driving cost down in parallel with performance, microbolometer manufacturers were able to address new applications beyond their traditional military markets. In the case of LiDAR manufacturers, they could keep performance high to keep their position into the robotic space. Or they could choose a lower performance approach with solid-state technology to serve the ADAS market, then increase performance gradually to disrupt the robotic space in a few years’ time. This is truly where the strategy game is happening right now. Some industrial players are trying to keep their monopolistic position with high performance, but are also investing in the low-end approach.
Do we really need LiDAR – and will the first movers be the winners?
Nothing in the LiDAR industry is simple, as demonstrated by the different potential grades in LiDAR specifications. These include high-resolution LiDAR for front detection, corner-LiDAR for edge scanning, rear-mounted LiDAR with low resolution and top LiDAR with 360° degree view and long range. There are 100 companies hunting in an undefined market, with moving targets in term of specifications, price, compatibility and sensor fusion, alongside well-entrenched camera modules and radar systems. Sensor fusion with multiple camera images and multiple radars is really complex. Add multiple LiDARs on top of that, with less than a fraction of a second of reaction time and you have a very complex computational problem.
There is no doubt that automotive LiDAR is progressing rapidly today, but alternative approaches exist. Tesla has announced that it will not use LiDAR for its autonomous functions, and TuSimple, a Chinese company, is developing autonomous trucks without LiDAR. Also, the recently-released Cadillac Super Cruise relies heavily on maps and existing sensors. On the technology side, automotive radar is beginning to gain imaging capabilities thanks to beam steering. Vision processing and artificial intelligence are helping cameras to better detect and classify objects and therefore understand three-dimensional space. Infrared cameras might also play a role for night vision. All of these technologies are improving at an impressive rate, creating a tough competitive space for ADAS LiDAR players. As such, it is too early to determine which technologies will feature in autonomous cars 15 years from now. However, we’re certain that today redundancy and complementarity between sensors is mandatory for ADAS cars and robotic vehicles alike to reach their potential. Some stealth players are mastering all the pieces of the puzzle, and could clearly have advantages over others. Players well positioned in cameras, radar, ultrasonic and LiDAR will clearly have strategic advantages.
So what could happen?
In the end, the winners of today may not be the winners of tomorrow. But whoever wins, overall, robotic or ADAS cars need to be perceived as safe, optimized and cost effective. As always in the automotive industry, these markets will be highly segmented. High-end car manufacturers will deliver high-end safety features with redundancies and new functionalities. Low-end car makers will focus on entry level functions. This is exactly what happens for safety functions in existing cars. LiDAR will certainly be a strong market for autonomous driving and robotic cars – but only in certain market segments. To learn more, please see the new report released by Yole Développement LiDARs for Automotive and Industrial Applications 2018.
Dr. Guillaume Girardin works as Director, Photonics, Sensing and Display at Yole Développement (Yole), the «More than Moore» market research and strategy consulting company.
Guillaume holds a Ph.D. In Physics and Nanotechnology from Claude Bernard University Lyon 1 and a M.Sc. in Technology and Innovation Management from EM Lyon School of Business.
Will automotive change the LiDAR market? – Get more
Sensors for Robotic Vehicles 2018
High end industrial sensors will win in the emerging robotic vehicle industry.
Get more here.
LiDAR for Automotive – Patent Landscape Analysis
High end industrial sensors will win in the emerging robotic vehicle industry.
Get more here.
Get ahead of the robotic vehicle wave and get to know the key technologies for this transportation revolution to happen.
Yole Developpement and KnowMade have organized a special webcast on April 17th 2018 – you can still register here and get access to the replay, do not miss it!
After reviewing the key insights of our latest report Sensors for Robotic Vehicle 2018 they had presented two special focuses: the first one on computing technology for AV and the second one on the LiDAR IP landscape. The complete presentation part last 30 minutes and is followed by a 30 minutes Q&A with the analysts.
Related Reports and Monitors
Melexis Far Infrared Thermal Sensor MLX90640
Reverse Costing - Structural, Process & Cost Report
CMOS Image Sensor Service – Imaging Research