Many insiders expect significant consolidation in the LiDAR market in 2019, and many suppliers of the technology are jockeying for position in the rationalizing market. Ouster is one of the companies hoping to survive and thrive, saying it is showing momentum in the business at just the right time, hoping to solidify its position as a leader in the market believing it has the most promising technology for the long term.
After announcing $27 million in Series A capital in December 2017, Ouster has worked with existing investors and new backers to secure an additional $60 million in equity and debt funding. It includes investments from Runway Growth Capital and Silicon Valley Bank, as well as additional funding from Series A participants Cox Enterprises, Constellation Tech Ventures, Fontinalis Partners, and Carthona. The additional investment has helped to develop Ouster’s product lines, including the launch of the OS-1 128 sensor, and fund the expansion of its production facilities.
It recently crossed the 400 customer mark from 15 different industries, and while most of the best names are under strict NDA it says the diversity in its customer base is worth noting. While most LiDAR companies are focused on automotive, especially OEM consumer vehicles that are many years away, company execs are taking a broader approach. They say that Waymo selling to robotics companies seems to be a validation of this strategy. Ouster sells to companies in the robotics, drones, mapping, defense, building security, mining, and agriculture industries.
The big operational news from the company is the late-March opening of a new manufacturing center at its headquarters in San Francisco. The center will include space for nearly a hundred manufacturing staff, reliability and quality assurance equipment, and sensor calibration that is included with every sensor. It will have the capacity to assemble, calibrate, and ship several thousand sensors, or $25-50 million in inventory, per month toward the end of 2019.
All of Ouster's sensors are based on the same multi-beam flash technology that has given the company significant momentum in the competitive LiDAR market. “In fact, very little besides the chips inside change from one model to the next, making our sensors much easier to manufacture than others since they can be made on the same production line,” said Angus Pacala, Ouster CEO.
Of the company’s approach, “the key difference is the all-semiconductor design, which is a dramatically simpler architecture than any other high-resolution sensor,” added Pacala. “Without sacrificing on performance, you get a more durable, smaller, lighter, and less expensive 3D LiDAR.”
One of the more unique aspects of Ouster’s sensors is their 850-nm operating wavelength. The lasers in a LiDAR sensor must overcome ambient sunlight to see obstacles, and industry engineers often choose operating wavelengths in regions of low solar flux to ease system design. The company’s decision to operate at 850 nm runs counter to this trend, wrote Pacala in a November 2018 blog post.
A plot of solar photon flux versus wavelength at ground level shows that, at 850 nm, there is almost two times more sunlight than at 905 nm, up to ten times more than at 940 nm, and up to three times more than 1550 nm—all operating wavelengths of legacy LiDAR systems.
“We’ve gotten plenty of strange looks for our choice given that it runs counter to the rest of the industry,” wrote Pacala. “However, one of our patented breakthroughs is exceptional ambient light rejection which makes the effective ambient flux that our sensor sees far lower than the effective flux of other lidar sensors at other wavelengths, even accounting for the differences in solar spectrum.”
The company claims its intellectual property turns what would ordinarily be a disadvantage into a number of advantages: better performance in humidity; improved sensitivity in CMOS (); high quality ambient imagery; and access to lower power, higher efficiency technologies.
“Ouster’s technology is focused on doing more with less - less cost, less power, smaller size, and fewer components,” wrote Pacala.
Some customers are surprised that Ouster can achieve such high performance with an all semiconductor approach, but Pacala says it is because of a unique laser-detector combination: vertical cavity surface emitting lasers (VCSELs) and single photon avalanche diodes (SPADs). VCSELs and SPADs are technologies broadly deployed in markets outside of LiDAR (in smartphones, optical mice, medical equipment, and telecom infrastructure), but until now there have been insurmountable challenges to using them in high-resolution LiDAR systems.
“As far as we are aware, Ouster is the first company to not just demonstrate but also commercialize a combination of long-range and high-resolution sensing with VCSELs and SPADs in all ambient conditions,” wrote Pacala. “There’s a reason for this: it’s super hard… Ouster is the first company to commercialize a high performance SPAD and VCSEL approach.”
For three years, company engineers have been developing the multi-beam flash technology to overcome the challenges that kept single-chip technologies out of LiDAR previously, culminating in the multi-beam flash architecture that plays to the strengths of VCSELs and SPADs. In this case, its “flash” approach refers to the idea that every pixel in the sensor is illuminated by the laser and actively collecting light simultaneously like a camera with a flash, and “multi-beam” refers to the fact that the scene is illuminated with precision beams of light instead of a flood.
The precision-beam approach goes back to Ouster’s focus on efficiency, wrote Pacala: “The flood illumination in a conventional flash lidar, while simpler to develop, wastes laser power on locations the detectors aren’t looking. By sending out precision beams only where our detectors are looking, we achieve a major efficiency improvement over a conventional flash lidar. Though VCSEL’s aren’t as bright (yet) and SPAD’s aren’t as efficient (yet) as the high cost legacy technologies currently employed, but what each lacks in raw performance they make up for in virtually every other metric: reliability, durability, low noise, high temperature operation, electrical efficiency, compactness, cost, direct integration with peripheral components, massive peripheral R&D activity driven by the consumer electronics industry, and equally massive room for fundamental performance improvements.”
Since Ouster launched in late 2017, the company has announced four lidar sensors with resolutions from 16 to 128 channels, as well as two product lines: the OS-1 and OS-2. The OS-1 is a mid-range sensor, whereas the OS-2 is a long-range sensor.
“The OS-1 is currently being used across at least 15 industries in dozens of different use cases,” said Pacala. “We expect the OS-2 to primarily be used for robotaxi and mobility applications, but may find a fit in other applications as well, such as industrial robotics and drone surveying.”
What started as a four-person team working in a tiny warehouse three years ago has grown to over 100 full-time employees across engineering, operations, business development, and marketing. The company expects to nearly double its headcount in the coming year to support further product line development and meet the global demand for its sensors.
The latest of those sensors is the OS-1-128. Launched in early 2019 and shown at CES 2019, the company says it is the highest resolution LiDAR on the market with 128 individual laser beams. And it comes with no change in size, mass, power consumption, or ruggedness compared to the company’s current OS-1-64 unit.
The OS-1-128 has a 45° vertical field of view, the widest available of any commercially sold high-performance LiDAR sensor. It produces 2.62 million points per second to form the dense point clouds necessary for object detection and machine learning in autonomous driving.
In 2018, Ouster announced its unique ability to achieve correlated 3D and 2D range, intensity, and ambient camera-like images with its current high-resolution OS-1-64, which is sold into a variety of markets including autonomous vehicles, robotics, drones, mapping, smart cities, building security, and industrials.
The OS-1-128 takes that imagery to the next level with double the vertical resolution, further accelerating applications of 2D deep learning algorithms on 3D point cloud data. The company expects the new sensor to be available for volume purchases in the summer of 2019.