Orlando's self-driving startup, Luminar Technologies
Can an Orlando startup overcome a host of challenges to help Central Florida make a name in self-driving autos?
In the global race to develop self-driving cars, Florida took an early lead. Eager to draw a new wave of tech, the state enacted legislation to encourage vehicle trials. Out-of-state automakers, trucking firms and transit tech companies drove in.
The state succeeded at attracting road tests of self-driving vehicles, but lacked home-grown firms engaged in developing the technology that’s involved — until 2017, when an Orlando startup, Luminar Technologies, emerged with gear and software that enable self-driving vehicles to “see” and understand conditions on the road.
For cars to drive themselves, they need to perceive and analyze their surroundings, from road signs to pedestrians to lane indicators and other vehicles. It’s a complex puzzle that most think will take a variety of sensors to solve.
Luminar specializes in a piece of the puzzle that’s increasingly seen as critical — lidar — light detection and ranging or, even simpler, laser radar. Research firm Global Market Insights forecasts the automotive lidar market will reach $3 billion in revenue by 2024. Some 80-plus firms are chasing that market.
With early endorsements from Volvo, Toyota and Volkswagen-Audi, Luminar is considered one of the firms that could survive as automakers settle on the technology they’ll use in self-driving vehicles.
The stakes are big — not just for the company, but for Central Florida and its efforts to beef up the non-tourism sectors of its economy.
“We are giving companies like Waymo and other large companies a run for their money,” says Jason Eichenholz, Luminar’s co-founder and chief technology officer. “Florida’s a really great place to be doing what we’re doing. The talent that’s here locally is second to none.”
That talent pool owes to a quirk of biography. Beginning in the late 1960s, an entrepreneur and laser expert named William C. Schwartz began founding laser companies in Orlando. Schwartz is generally regarded “as the father of lasers and optics in Orlando” and of UCF’s College of Optics and Photonics, says University of South Florida physics professor emeritus Dennis Killinger. The Metro Orlando Economic Development Commission’s innovation awards are named for Schwartz, who died in 2000 at age 73.
Among his endeavors, Schwartz funded the first research project undertaken by Eichenholz in 1993 as a young UCF graduate student. A Brooklyn native raised in New England, Eichenholz as a kid performed laser experiments in his basement and attended a National Science Foundation program in high school. UCF’s reputation in optics drew him to Florida, where he earned a master’s and doctorate in optical sciences and engineering.
Embarking on a career in the field from Orlando, he was introduced a few years ago to a California couple who wanted a mentor for their precocious son, Austin Russell.
Russell is said to have memorized chemistry’s periodic table at age 2. By his teens, Russell was doing laser experiments in his garage and already had applied for his first patent. Eichenholz, remembering his own youth spent doing laser experiments, took Russell under his wing. They went to trade shows and conferences together
In 2012, Russell, then 17, founded Luminar in California, with Eichenholz joining shortly thereafter as co-founder and chief technology officer in Orlando. Russell went off to Stanford University the next year but left after three months to accept a Thiel Fellowship — a $100,000 offer from PayPal founder Peter Thiel to young entrepreneurs to skip college in favor of pursuing a promising idea. In Russell’s case, it was lidar.
Lidar has been at the forefront of self-driving developers’ minds since the Defense Advanced Research Projects Agency’s self-driving vehicle “Urban Challenge” in 2007. Most vehicles that finished the DARPA challenge used lidar from Velodyne, a California-based company.
As automakers began looking to massproduce self-driving vehicles, they saw that no single type of sensor solved all issues. Cameras are necessary but see only in 2-D and don’t readily measure distance. Radar sees far but struggles with other tasks.
Lidar “sees” in 3-D, meanwhile, and functions just as well in both darkness and daylight. It calculates distance and detects motion effectively. But adapting lidar for the mass auto market presents plenty of challenges — challenges so daunting that some players don’t use lidar at all. Elon Musk’s Tesla says cameras are sufficient. San Francisco-based Starsky Robotics, which has operated a tractor-trailer rig with no humans on board in trials on Florida’s Turnpike and Tampa’s Selmon Expressway, says lidar isn’t necessary for its model. Starsky’s model is to have humans using cameras remotely pilot trucks onto highways. Starsky then lets the technology do the driving before humans, via camera, take control again to steer trucks onto secondary roads.
Among Lidar’s difficulties: Lidar can’t tell a green light from a red one, as cameras can. There also are issues in lidar with power consumption, cooling, getting unit prices down from tens of thousands to hundreds of dollars and making devices small enough to blend into a vehicle’s design. Lidar also must be proven durable. Car parts have to be able to withstand extreme cold and extreme heat and constant vibration.
While some companies have sought to solve those problems by adapting off-the-shelf tech for use in cars, Russell and Eichenholz have designed everything from scratch for the auto market. Russell dismisses the other approach as “Frankenstein solutions.”
Silicon Valley culture
Luminar now has a portfolio of nearly 50 patents and has applications pending for three times that many. Luminar employs 250 in its three buildings at the Central Florida Research Park who are developing the hardware sensors: Lidar’s “eyes.” Another 100 are in California, of whom 60 work on the software that understands what the sensors see. Automakers then will need to integrate it with their own software that makes driving decisions about braking, turning and accelerating.
“We make that job much, much easier,” Eichenholz says. With Luminar’s tech, cars “are no longer nearsighted.”
The company claims Orlando as its headquarters, although the CEO and many top executives work in California.
Luminar imported Silicon Valley culture to Orlando. It has stock options for employees and free snacks and beverages, including cold-brew coffee, at all times. A chef prepares meals daily. On the Glassdoor website, employees in 25 reviews gave Luminar mostly positive marks, with praise for the tech but with some complaints about disorganization, yelling in meetings, long hours and poor human relations.
In March, Rhonda Newcomb, who worked at the company in 2017 and 2018 in a tech capacity, sued Luminar alleging overtime pay violations and sexual harassment. In her suit, she says she regularly worked 45 to 60 hours per week without getting time and a half. After she ended a brief social relationship with one manager, she says, he continued making “inappropriate sexual overtures.” She says she was sexually harassed both verbally and physically by another manager and that a co-worker exposed himself to her and others and refused to cooperate with her on work when she rebuffed his advances. She says her complaints were ignored and she was put on a performance improvement plan by one of the supervisors she alleges harassed her. She was fired in April 2018.
Luminar’s attorneys have moved to dismiss the case. The company says, “There is no place for harassment at Luminar, and we are committed to strictly enforcing our anti-harassment and antiretaliation policies. Luminar denies and is defending against Ms. Newcomb’s allegations and cannot comment further on pending litigation.”
Luminar faces other challenges. More than 80 companies, according to Autonomous Vehicle Engineering magazine, are developing lidar systems for the autonomous vehicle market. No one expects more than a half-dozen to survive. It’s expected that most of the best will be acquired by automakers or first-tier suppliers to the auto industry. Luminar is in the top half-dozen lidar contenders, says Sam Abuelsamid, the principal analyst covering autonomous vehicles at consulting firm Navigant. But he says Velodyne, the company whose lidar dominated the DARPA challenge more than a decade ago, is “far and away the biggest.”
The arguments for each company’s competing approaches get highly technical. Luminar, for example, and others champion lidar that operates in the 1550 nanometer wavelength, which they say works better and is safer for the human eye than Velodyne’s 905 nm. Velodyne says its 905 lidar is superior.
“Everyone’s going to make the argument that defends the technology they’re developing,” Abuelsamid says. Each hopes to win access to automakers’ test fleets and emerge with their gear as the standard an automaker adopts.
Luminar has received early affirmation. Volvo’s investment arm — the Volvo Cars Tech Fund — last year made Luminar its first investment. Toyota has said Luminar’s “level of data fidelity and range is unlike anything we’ve seen.” The self-driving development arm of Volkswagen and Audi similarly praised Luminar’s tech. They are the only three Luminar has publicly identified of the 12 out of 15 top automakers with which it says it is working.
But relationships aren’t exclusive. “While we have been public about using Luminar technology, our team is constantly exploring new breakthrough technology from other suppliers,” says Rick Bourgoise, Toyota Research Institute’s communications manager.
The ground shifts rapidly. In December, Audi-Volkswagen’s autonomous development unit announced a partnership with Luminar and called Luminar’s technology “clearly above the pack.” Four months later, the unit called rival California-based Aeva’s lidar solution the “clear top choice.”
In July, Luminar looked to have made a breakthrough. It announced its technology, dubbed “Iris,” as the solution to all the problems adapting lidar for self-driving autos. Luminar said Iris would be ready to debut on production lines in 2022 and meets all the performance, durability, safety and cost requirements for the top levels of autonomous driving. (Autonomous driving specifications range across a spectrum from simple things like cruise control to cars that don’t need a driver.)
Luminar said Iris provides cameralike resolution and radar-like range out to 250 meters for under $1,000 and in a package small enough to be integrated into a car body’s design. Iris is the “highest performing system out there on the market,” Eichenholz says. (Luminar also announced a less expensive version of Iris for the driver-assisted market — lower down the autonomy scale.)
But the same day Luminar announced Iris, Volkswagen — whose autonomy unit was working with Luminar — announced that it will take an ownership stake in Ford-backed Argo AI, which has its own lidar team, and Volkswagen said it also would fold its autonomous vehicle unit into Argo. Luminar says it’s still working with the unit.
The Ford-Volkswagen deal highlighted the consolidation roiling the field and the money at play. Luminar, with $250 million in capital raised and a valuation of $900 million, is small relative to others, albeit focused narrowly on lidar, while others are more comprehensive. Volkswagen invested $1 billion in Argo AI and valued the unit it folded into it at another $1.6 billion. Honda Motor put $2.75 billion into San Francisco-based Cruise, a General Motors subsidiary. Cruise is valued at $19 billion.
Meanwhile, the auto industry is coming to a reckoning that truly self-driving cars are a way off. In but one example, Cruise recently said it was abandoning plans to launch a self-driving taxi service this year and wouldn’t commit to do it in 2020. The sensors and the artificial intelligence that control the vehicle simply aren’t up to the task of full autonomy. Three high-profile fatalities have occurred in Teslas — including two in Florida — operating in auto-pilot mode, and last year a fatal collision occurred in Arizona between a pedestrian and an Uber test vehicle.
The difficulties automakers have encountered have limited demand for lidar products from Luminar and other makers. In 2017, Luminar said it would produce 10,000 lidar units that year in Orlando. The company now says only that it’s ramping toward that number. Eichenholz says automaker test fleets have shrunk from hundreds of units deployed to as few as 10, cutting demand for sensors dramatically.
Abuelsamid, the Navigant principal analyst, expects only low production in the next couple years. By 2022, the firm forecasts that only 300,000 vehicles out of 95 million manufactured globally will be able to truly drive themselves.
Eichenholz says that without Luminar’s tech, it will take the industry 20 years to have masses of self-driving vehicles on the road. With it, he says, automakers can shave a decade off development.
“I can assure you this is a much harder problem than people are admitting, and the industry is a little bit out over its skis right now,” Eichenholz says. “What we’re witnessing right now is the single largest transformation to transportation since the Model T. It’s like the difference from going from horses to horseless carriages.”
At the Automated Vehicles Symposium conference, held in July in Orlando, St. Petersburg Republican state Sen. Jeff Brandes, who has spearheaded Florida’s autonomous vehicle-friendly laws, highlighted Luminar and talked up the region’s lidar talent. The Legislature made Florida the second state to permit self-driving trials and later passed a law allowing fully autonomous vehicles on the roads beyond testing and also truck platooning, in which trucks line up in a convoy with their speed controlled in unison. Brandes’ legislation also mandates Florida governments consider autonomous vehicles in their road planning. Florida is also notable for what it doesn’t require — a permit for testing as in California or a police escort as in New York.
The state’s laws — described as “more lax than some other states,” by Automotive News — have helped make Florida a national leader in trying to entice companies to test self-driving vehicles here. As a result, big names like Ford and Volkswagen and lesser-known manufacturers have come here for trials. Voyage, a self-driving taxi company based in Silicon Valley, has its self-driving vehicles in The Villages. Transdev, a French public transport company that makes autonomous shuttle buses, operates autonomous shuttles at Babcock Ranch near Fort Myers.
The state passed additional legislation this year that, among other provisions, permits self-driving ride-hailing companies. The new law eliminated the need for self-driving auto operators to have a driver’s license and, in fact, declares the system is the “operator” rather than a person. The state also exempts autonomous vehicles from the normal legal obligation of drivers involved in a crash to provide police and others involved with their personal contact and license information and vehicle registration and to render aid in the event of an injury — as long as the actual vehicle owner, or the vehicle itself, notifies police.
In addition to the legislation, the state also has funded the $42-million SunTrax facility in Auburndale, which tests electronic toll collection technology, and, by 2021, an autonomous-vehicle testing track. The track, near Florida Polytechnic University, will be operated as a partnership between Florida’s Turnpike Enterprise and the school.
Read more in our September issue.
Select from the following options:
* offer valid for new subscribers only
Measles is a 'heat-seeking missile' experts warn as Florida outbreak grows
Florida lawmakers continue debate over school red tape, education spending and more