Richard Llewelyn published the fourth and last novel in his series about life in the mining communities of the South Wales valleys in 1975, eight years before his death. Lidar was still mainly pointing upwards and he would not have known it – yet less than fifty years later it pervades our lives. As LIDAR Magazine returned to the Bay Area and Silicon Valley after a 16-month absence owing to the pandemic, the poignant title of the novel reverberated in your scribe’s mind. Lidar is prospering and it was a pleasure and relief to be back after so long away.
Before the visits, there were vacation days in San Francisco, during which we made a family visit to the Exploratorium, a magnificent, hands-on museum on the waterfront, to see a small exhibit about Waymo sponsored by Google. This shows what’s behind the housings on the Jaguar i-Pace and explains the differences between imagery and point clouds. Families were absorbed with lidar! Indeed, there seemed to be Waymo Jaguars everywhere in the city and the company recently announced a robotaxi service in San Francisco. They’ve also announced that they will no longer resell their Honeycomb lidar, restricting it to in-house use. Interestingly, I was in the process of getting into the left rear door of my son’s car when a Waymo Jaguar approached and I’m sure it saw me and made a small diversion. On the other hand, we were sitting at an outdoor bar, extended by one of the ubiquitous “parklets” rapidly constructed to furnish more outdoor seating. The bar was at a corner and a large delivery van had parked beside the parklet. A Waymo Jaguar came round the corner and I noticed that the driver was steering it past the van. Neither the van nor the parklet would have been in its map database, I surmise, so although the lidar sensors and software would have prevented a collision, the car would perhaps have been unsure of exactly where to go. Some think that the whole issue of how advanced AV software really is remains open[1]. I also noticed a number of Cruise autonomous vehicles – Nissan Leafs. General Motors acquired Cruise in 2016. Waymo, Cruise, DeepRoute and others have permits to run robotaxi services in California. Further firms have permits to operate AVs, some with drivers and some without. It’s happening!
There followed four visits, which will be reported in full in a long article later. Firstly, I sat in a large conference room in Sunnyvale with Quanergy Systems’ co-founder and chief development officer Dr. Tianyue Yu and marketing manager Sona Kim. Tianyue dampened my ardor right away by saying she couldn’t comment on details of the firm’s ongoing SPAC. Nevertheless, we had a fascinating conversation on where the firm is going. I asked Tianyue – as I asked her counterparts in competing companies – whether all the SPACs are like a rising tide and all the lidar companies, boats bobbing upwards upon it. Her preferred analogy in the full article. The pickup truck that rear-ended me later in the day on one of San Francisco’s precipitous hills was, sadly, ADAS-free.
In the evening, I dined with Becky Morton, an ASPRS veteran who is co-founder and CEO of Geowing Mapping, a UAV surveying and mapping company in Richmond, California. Her company was formed in 2015 and is prospering. It specializes in imagery, not just RGB but also multispectral and thermal. Geowing has an impressive list of clients, several of which recognize the company’s reliability and quality by placing repeat business. Becky and her team have followed the principle of not investing in lidar systems, mainly owing to the high cost and the risk to the hardware if there is a drone crash (yes, these happen, even to experienced operators). But they offer lidar services through partner companies, using both UAVs and crewed aircraft, as a result of which Geowing’s portfolio of services is broad. Indeed, Becky and her colleagues have gained considerable experience in planning lidar missions and processing the data into deliverables. Would she rethink her position as the price of lidar falls, partly as a result of the endeavors of the other firms visited on this trip? Yes, possibly.
The following morning I was in San Jose to see Cepton Technologies. I met with co-founder and CEO Dr. Jun Pei; VP, finance & strategy Hull Xu; and marketing & communications manager Faithy Li. Cepton had been my last visit for the magazine before the pandemic. We reported then on its move to a new headquarters, with the interior design closely supervised by Jun himself, but the company has moved again! The address is on West Trimble Road – the paths of geospatial excellence and automotive lidar seem invariably to cross. The new facility is palatial and aesthetically pleasing, partly because it was previous occupied by AMSL, which departed for a new campus. Hull’s name is pronounced like the declining fishing port in eastern England, but he’s on the up and up, an ex-banker, absolutely in command of the SPAC formalities. Educated at Davis, Stanford and Berkeley, Hull continues the Stanford emphasis in the Cepton leadership team and, indeed, as I was leaving, we bumped into the new CFO, Dr. Winston Fu (MIT, Northwestern, Stanford). Jun emphasized the importance of contracts with automotive manufacturers or tier 1 suppliers – these are the key to being able to sell lidar sensors in the hundreds of thousands or beyond. This is a point also made in our article about Innoviz. Furthermore, the day I visited Cepton, Luminar Technologies, yet another SPAC-fueled lidar producer in the Valley, announced two tier 1 partnerships. Jun hinted at an imminent announcement! He acknowledged one disappointment: as Cepton grows, he no longer knows the first name of every employee – and he would like to.
Hanging on the wall of the reception area at Quanergy were several patents. On one of them was the name Angus Pacala, who was my next quarry. I rushed to downtown San Francisco and located Ouster in a labyrinthine but charming converted tee-shirt factory, complete with beautiful wooden beams. Ouster, named after a technologically advanced tribe from the Dan Simmons Hyperion series dedicated to using technology to live in harmony, was founded in 2015 by Angus and Mark Frichtl, who came from Quanergy, and two more classmates – guess which university. Angus works on a desk in an open area, the same as any other employee, but behind it he has a sort of telephone box (is it a Tardis?), into which he and his laptop retreat to make calls or seek silence. We had a wonderful conversation – Angus is young, unassuming and brilliant. He has a gift for pithy phrases summing up Ouster’s technology and mission, such as making lidar “affordable and ubiquitous”, “building the eyes of autonomy”, and “digital lidar for everything”. Our conversation will be recounted in depth in the article. After the discussion, I was given a super tour by product marketing manager Tom Grey and one of the topics we covered was software. As readers know, I have been trying to arrange the lidar suppliers on some sort of capability spectrum, from those that don’t do much beyond the sensors themselves, to those which offer considerable software, sometimes called a “perception layer” and often dependent on deep learning. Tom emphasized that Ouster has extremely strong software capabilities and a huge complement of software experts on its staff.
That’s enough appetizers – the full article will take some work as I try to assimilate all I learned and transcribe many hours of conversations. Certain trends on the lidar sensor side are clear. Many lidar companies have easier access to capital owing to SPACs. Huge automotive contracts are one way to success, though in 2021 and the next couple of years these are more likely on the ADAS than AV side – this of course will change as time passes. As a result, the companies also have their eyes on other markets. Geospatial is a small one, but others, such as smart cities, abut on our world. Sensors will continue to become cheaper, higher performance and more robust. The accompanying software will grow in capabilities and strength. All this will be available to the geospatial community, which will benefit indirectly from the vast size of the other verticals which the lidar suppliers are pursuing.
I completed this note while in Las Vegas for the Commercial UAV Expo, meeting people again after so many months remote. Geowing Mapping is a reminder that, with experience and considerable hard work, rich rewards are available from UAV-based geospatial services. Even companies with modest capital equipment budgets can consider lidar sensors as prices fall, but much is possible – as it is in the automotive lidar world – through partnerships up and down the supply chain.
So much learned, so much to ponder – it’s a privilege to be back! I’m so appreciative of the welcome and time that the folk I visited gave me and their patience as I asked my questions. We are part of a quite remarkable technology, business and professional community, oozing with talent, imagination and initiative!
[1] Anon, 2021. AI for vehicles: is it smarter than a seven-month-old, The Economist, 440(9261): 65-66, 4 September 2021.