Velodyne Preps for Volume Lidar
The $500 Velarray H800 is aimed at ADAS and automated production at scale.
Prior to 2020 there were allegedly 80 to 100 “lidar companies” in the world, depending on who in Silicon Valley was compiling the list. Such was the over-inflated bubble of autonomous mobility prior to the pandemic. How many of those fledglings retain any viability today is anyone’s guess, but one thing is certain: Velodyne is real, and remains the best-known player of the lidar world.
So, its announcement last November of the Velarray H800 – first in a new family of solid-state lidars aimed at automotive ADAS and AV applications – continues to resonate among OEMs and Tier-1 integrators. In development since 2017, the H800 is based on Velodyne’s MLA (micro-lidar array) technology that uses eight edge-emitting lasers. It also features the company’s proprietary photo detectors and its own ASIC microchip – an application-specific integrated circuit.
All of this makes the Velarray H800 a robust, capable and package-efficient sensor that can tuck into the rear-view mirror module behind the windscreen. The H800’s price – less than $500 per unit, the company claims – created a buzz, as would be expected for any new lidar. As a technology, lidar still struggles on the cost front; $500 represents a significant drop in cost from the $10,000 electro-mechanical lidars of yore but is still wildly expensive for any non-powertrain or driveline component.
Velodyne CEO Anand Gopalan put the H800’s price into perspective in an interview with SAE Media. “At CES this year we talked about a $100 Velobit lidar for ADAS – the starting point in a family of lidars that include the Velarray suite priced from south of $500 to maybe $1,000 for the high-end,” he said. “That’s a clear line of sight by us to be able to achieve that today. But we have plans to further reduce cost, as you’ll see with this first wave of new solid-state lidars.”
While Gopalan could not talk publicly about Velodyne’s technical innovations that will enable cost reductions, he asserted that “cost is not going to be an issue even for [SAE] Level 2. With our expanding range of products, we can place the technology for the specific application.”
Aimed at all SAE levels
Gopalan, an electronics and microsystems engineer, explained key features of the solid-state H800 and talked about the enablers that led to its development. “As the [solid-state] technology platform matured, we started working with some of our lead automotive OEM customers around building a scalable product that’s really suitable for their needs in ADAS, from [SAE] Level 2 all the way to Level 5 full autonomy. We expect it to be used mainly in Level 2 and Level 3 applications, of course. The H800 was really architected with automotive-grade feedback.”
The trend of adding more capability to Level 2 driver-assist systems prompted Velodyne’s development of the Velarray product. “Three years ago we were hearing from our OEM customers a realization, that lidar has to be part of the solution to create robust ADAS Level 2-Plus and Level 3 systems,” Gopalan said. He frequently is asked why vehicles need lidars when they have cameras and radar. Velodyne in 2020 conducted a test comparing what he termed “a state-of-the-art” camera/radar system with a Velarray-based system incorporating Velodyne software.
“We just took a simple function – pedestrian automatic emergency braking – but we wanted to test it in the nighttime,” he said. “When we did that, we saw that the state-of-art camera/radar-based system failed more often than not. But the Velarray was able to detect the pedestrian and apply the brakes every single time at night. That’s the power of lidar – it brings its own light to the party.”
Design for manufacturability (DFM) and scalability were key tenets of the Velarray program, according to Gopalan. “For mass adoption of lidar, the product must be producible in the hundreds of thousands, with multiple sources for the key components with automotive-sized manufacturing footprints,” he said.
As of early November 2020, the new H800 was in early production on Velodyne’s pilot line in San Jose. “The line is running at full capacity and we’re bringing up a line with our manufacturing partner Fabrinet in Thailand as well, to manufacture this at scale,” Gopalan said. Fabrinet is considered the leading contract manufacturer for electro-optical systems.
For automotive lidar to become ubiquitous, the cost to the OEM and the end price of the technology are critical. “Velodyne’s vision is to improve and democratize vehicle safety across the entire segment,” Gopalan said. “That’s the reason we’re working hard to provide affordable ADAS technology to the marketplace.”
Tech analyst’s take
Phil Amsrud, senior principal analyst, Semiconductors, IHS Markit, spoke recently with SAE Media about the trend toward solid-state lidar technology and the industry’s need for lower-cost systems. Here are three key insights:
- The buzz about solid-state: “A few years ago there was a lot of talk about the importance of getting to solid state because of its inherent reliability versus the electro-mechanical lidars and versus MEMs-based lidar (micro-electromechanical). The automakers want good performance at a fair price for their vehicle applications, whether it’s solid state or not.”
- $500 per unit is still extremely costly: “$500 lidars are a positive step for the industry in terms of cost reduction, but they’re not going to make a big dent in the L2 space. Those lidars are probably going to be more for the few L4 systems released for robotaxis and mobility-as-a-service solutions. “It’s a price/performance discussion – how many degrees of view will $500 buy me? How much range? And what is the application? Not every lidar needs to deliver 200 meters of performance. There is value in short-range, lower-cost lidars. And, if a more expensive unit’s performance means I don’t have to put multiple lidars on my vehicle, it can justify the price. Also, how does that price compare to competing technologies like imaging radar sensors? A robotaxi or people-mover operating in a geo-fenced area is a very different use case than driving in your own car from Manhattan to Los Angeles on a variety of roads.”
- The ever-evolving lidar market: “I see the lidar market offering mixed technology based on the use case and the value proposition. Right now, it’s hard to say which of the various companies and business models will turn out to be dominant. Captive manufacturing allows you to control quality and everything necessary to make a precision solution. And in this segment of automotive, there is something to be said for controlling your cost until the market gets established.”