At B.1.617.2’s Mercy

We had hoped by now to have returned from the pandemic to normal life and the hot summer. Delta, however, has scuppered that and we are battening down the hatches again. There are some face-to-face and hybrid events, but it takes resolution not to become depressed. Wimbledon was OK, but the Tokyo Olympics seemed rather sad. Usually I write enthusiastically about the top-quality online events that have been a wonderful substitute for the real thing, but my memory of one of those is the most depressing of all. The Esri User Conference took place remotely in July and was as beautifully orchestrated as always. I’ve commented on our website about the remarkable financial performance that Esri was able to report despite the drawbacks—the company’s success is simply indisputable. After quoting E O Wilson, Jack ended the opening plenary with the words, “It’s late in the day, but it’s not dark yet.” Finally, experts are accepting that our planet is probably slipping through our fingers. Billionaires bopping just above the Kármán line isn’t quite the solution we had in mind … and COP26 in Glasgow in November 2021 may be the last train, though the city in the winter may not be ideal for nurturing serious debtates about heat and drought. On the cover of its Heights 2021 supplement, xyHt asks, “Can airborne lidar save the planet?”, but the article being referenced is about making a digital twin of Earth¹, which is a tool but not a solution. We, the geospatial community, can help manage the fight against the effects of environmental degradation, for example managing wildfires, and we can provide base data for designing solutions, for example reforestation and the maintenance of infrastructure—but we cannot on our own create the will, on the part of the citizenry of the whole planet, to change.

While writing this, I had to dash to Home Depot for an infusion of tradescantia pallida. At one of the excessively numerous stop lights in our city, the pedestrian light was on and a lady with her dog passed in front of me. They arrived safely on the other side with 21 seconds still to elapse. The wait seemed interminable and, of course, there were several seconds more, after the clock had ticked down, before the lights themselves changed in my favor. How could the programming of the lights be improved, while safety for pedestrians, even those who make mistakes while in mid-crossing, is maintained? This very question arose in a webinar organized by Seoul Robotics, “Is there a blueprint for smart cities?” The speakers were William Muller (VP BD, Seoul Robotics), Richard Hardiman (CEO, RanMarine Technology) and Kevin Komstock (Director, City of Chattanooga). Stop lights came up many times in the conversation and the City of Chattanooga turns out to be adventurous, forward-looking yet responsible to its citizenry as it adopts new technology. Seoul Robotics’s CEO, HanBin Lee, has already been interviewed by the magazine² and we’ll be speaking with him again soon.

Shortly before the webinar, a Seoul Robotics press release on a new perception software product, Voyage, included the following argument, “The lidar market, which is on track to reach more than $3 billion by 2025, has become crowded over the past several years as the technology became synonymous with autonomous vehicles. The marketplace is flooded with lidar companies producing sensors to fuel the demand of this industry, but most sensors on the market are sold without any intelligence.” Users of Voyage include BMW, Mercedes-Benz and—wait for it—Chattanooga Department of Transportation. I noted the Vision Spectra conference on our website recently and I’ve been listening to some of the presentations. Artificial intelligence, usually in the form of deep learning, is becoming indispensable in machine vision.

Those of us on the geospatial side may be tiring of the financial news that often overshadows technology breakthroughs in the lidar world at present. It’s all about SPACs (special-purpose acquisition companies), which provide a route to a lidar company going public without a traditional IPO, while obtaining access to capital less arduously than would have been the case on the conventional venture capital route. It seems, however, that the private equity people, who’ve typically been less generous in valuing target companies than SPACs, may be bouncing back with their own SPAC vehicles³. For those who inhabit this world, it’s wildly exciting! We’ve seen SPAC deals with Aeva, AEye, Innoviz, Luminar Technologies, Ouster, Quanergy Systems and Velodyne Lidar in the last year; while this editorial was being written Cepton Technologies SPAC deal was finalized. We’ve already spoken to Velodyne⁴ and in this issue we have an interview with Innoviz co-founder and CBO Oren Rosenzweig.

Despite these deals hitting the headlines, technology has not hidden its light under a bushel. Quanergy is mounting a charge based on its optical phased array technology, reported in this issue by co-founder and CTO Dr. Tianyue Yu. This really makes a difference and the firm expects soon to announce an increase in range to 200 m. Meanwhile, Quanergy realizes that the autonomous vehicle is probably some years from widespread adoption, so in the shorter term it is diversifying markets. While automotive remains prime, the firm is pursuing what it calls the “internet of things market”, which includes verticals that Seoul Robotics has been talking about. Quanergy will use some of its SPAC-related capital to invest in sales as well as R&D. We’re trying to understand why one firm’s SPAC is better than another’s, but even if they’re all rather similar, easier access to capital will guarantee a rising tide and our geospatial world will be blessed with more robust, precise sensors at price points that would have been unimaginable even three years ago.

We also have an article by Dr. Gleb Axelrod, co-founder and CTO of Lumotive. I don’t have the electronic nous to differentiate competently between this and Dr. Yu’s piece. Suffice it to say that, in pursuit of miniaturization of lidar sensors, Lumotive seems to be following some similar paths. The demand is there. Just as I finished writing this, an article about robots for loading and unloading trucks arrived in my mailbox⁵. It described a robot arm as “festooned with sensors”.

In contrast to these firms, which are primarily producers of sensors and related software, Neural Propulsion Systems (NPS), another Silicon Valley start-up, has developed a complete system, NPS 500, designed for the automotive market, including its own lidar, radar and other sensors, integrated, packaged and accompanied by software. We include both an article about the NPS 500 and an interview with co-founder and CEO Dr. Behrooz Rezvani.

Thus the lidar suppliers from the automotive world have other markets in their sights, while their newly deep pockets facilitate R&D in both the sensors themselves and the necessary software. The geospatial world will surely benefit, while the specialist, “traditional” lidar suppliers will continue to hold sway for both low-noise, precise data from UAVs and MMSs as well as lidar acquisition from crewed aircraft and helicopters. In a very welcome development, Teledyne Optech have made several appealing webinars available on demand at teledyneoptech.com/en/events/webinars/. I’m going to make time for “Introducing the game changer: new AI software for efficiencies in lidar noise classification”. The whole lidar ecosystem is moving forwards at speed. Is it too speculative to wonder whether, with its experience of this measurement technology and the related data processing, geospatial firms can find a special role in the new but non-automotive applications, such as queue management at airports, stoplight control at intersections, and automation of ports, becoming involved in design, installation, data management and data analysis?

Indeed, we have two fascinating geospatial articles in this issue. The first is from Moncton, New Brunswick, Canada. The city needed a quick way to monitor vegetation above streets and sidewalks, making use of its existing staff and vehicles. It reached out to the local Trimble distributor, which proved resourceful and imaginative. The resulting low-cost, easy-to-use system is reinforcing Moncton’s twenty-first century civic pride. The other is from photogrammetrist Riadh Munjy, a doyen for decades at California State University Fresno, his student Jacob Lopez, and Chris Thornton from Caltrans. They report on practical tests, based on two very precise test fields, of UAV-lidar systems. This is a fine complement to the recent, popular article by Mark Meade and Kyle King⁶. Properly executed UAV-lidar missions provide accuracies high enough for almost all purposes⁷. We’ve seen other evidence of this in the many beautifully prepared webinars by Lewis Graham and Martin Flood of GeoCue. Geospatial expertise combined with the availability of first-class platforms and sensors at reasonable prices enriches our profession as we delight our customers.

Last but not least, we’ve included an “exit interview” of sorts with Lisa Murray, well known to many as the former geospatial lead at Diversified Communications, the Maine-based group responsible for Commercial UAV Expo, the International LIDAR Mapping Forum and more. As of January, Lisa has moved on to a new career in counseling but she gracefully took time to share insights gleaned from years spent developing collaboration amongst geospatial professionals. I must note that we collaborate closely with Lisa’s former group on our LIDAR Leaders Award Program, soon to be announced in its third iteration; Lisa was a driving force in the development of these awards.

We conclude on a light note. Recently I finished the third volume of Henry Kissinger’s memoirs of his years in the White House. This tome was centered on his service to President Ford, but of course there were inevitable references back to the Nixon years. Who should pop up but Donald Rumsfeld. All of us have, perhaps unkindly, enjoyed the craic wherein one matches his known knowns, known unknowns, unknown knowns and unknown unknowns to the constants, parameters, systematic errors and random errors of the mathematical model underlying least-squares estimation. Kissinger himself, however, penned the following gem, after an unsatisfactory briefing to the National Security Council by Director of Central Intelligence William Colby about the impounding of the SS Mayaguez by the Cambodians in 1975: “Some confusion was expected; totally inaccurate precision is harder to explain—indeed, I do not understand to this day what the basis for it was.”⁸ This is another gift to geospatial raconteurs and we can only wonder whether there could be some coup de grâce still to be divulged. Perhaps John Dean will reveal something about deep learning…