PARC Spinout Metawave Adds A.I. to Metamaterials for Autonomous Cars
A serial entrepreneur in the field of metamaterials is building a new company to develop the technology for use in autonomous vehicles and wireless communications.
Maha Achour co-founded Metawave early this year, spinning the company out of the storied PARC laboratories, a unit of Xerox, in Palo Alto, CA, with an exclusive license to commercialize metamaterials radar and antennas in two huge potential applications. In April, the company closed a “substantially large” seed funding round, says CEO Achour, declining to specify the amount, from investors including Khosla Ventures, Motus Ventures, Thyra Global Management, SAIC Motor, and Autotech Ventures.
Metamaterials represent a breakthrough in electromagnetic physics that is being used in new antennas that promise improvements in several dimensions over technologies they would replace. They consist of tiny, software-controlled engineered structures—often on a printed circuit board—that can form and steer electromagnetic beams in ways that were previously only possible with much larger and more expensive systems.
Metamaterials radars, for example, can perform the work of an advanced active radar system in a package sized to fit on a car or a drone.
Metawave, based in PARC’s Palo Alto headquarters, sees the technology as a key to unlocking autonomous vehicles.
“Radar is basically the key sensor of the car in the long-range field of view,” Achour says.
Other sensors, such as cameras from companies like Mobileye—which Intel is acquiring for $15.3 billion—and LiDAR (light detection and ranging), have shortcomings that radar does not. They have limited range and cannot “see” in adverse weather conditions, such as heavy snow or fog, Achour says.
However, the competition for metamaterials radars is not these other sensors, she says, but other, less-capable radar systems. Digital beam-forming radars, for example, require computationally intensive and time-consuming work to form, switch, and steer beams, particularly for sensing targets far away, Achour says. That means limitations for autonomous cars using the technology to see the road ahead.
“If you cannot see at 250 meters clearly over a wide field of view, you cannot drive very fast,” Achour says.
In addition to sensing, Metawave sees applications in vehicle-to-vehicle and vehicle-to-infrastructure communications that would be part of a future of autonomous vehicles at a large scale.
Xconomy has covered a quartet of well-funded metamaterials startups formed over the last several years by Bellevue, WA-based Intellectual Ventures, which acquired a trove of metamaterials IP from Duke University, Imperial College, and UC San Diego. They have attracted more than $290 million in venture capital funding from the likes of Bill Gates and are pursuing commercial applications in different fields. Metawave would compete most directly with metamaterials radar maker Echodyne, which is building technology for autonomous vehicles including drones and cars, and Pivotal Commware, which emerged earlier this summer to sell another version of the technology to the wireless communications industry, particularly as it eyes the transition to 5G networks.
Achour declined to provide much detail on her company’s technology and the intellectual property underlying it, citing the competitive landscape. “All that I can say is our solution is fundamentally different from what has been filed many, many years ago,” she says.
Metawave co-founder and CTO Bernard Casse filled in a bit more detail, saying Metawave uses “an all-electronic architecture.”
“We believe it’s faster, it’s less computationally intensive, and it’s more amenable to mass production,” he adds.
An interesting strategy point: Metawave is not just making metamaterials antennas. It is building a full radar package that it hopes to supply directly to car makers as they strive for fully autonomous vehicles.
“This is where the A.I. plays a big role,” Achour says, adding that Metawave is in discussions with a wide array of automotive industry partners.
Casse says Metawave is leveraging the rapid rate of progress in A.I. technologies to develop a series of algorithms for optimization, range-finding, and more. “What we want in the end… is a radar that’s smart, that can tell you there’s a child in your parking zone, that can discriminate between an SUV and a sedan, that can anticipate an accident by taking a look at traffic patterns,” he says.
Achour holds a PhD in physics from MIT, has more than 100 patents in metamaterials and other radio frequency technologies. In 2006, she was the founder and CEO of a company called Rayspan, which was the first to commercialize metamaterials technology—a passive version of it, rather than the more capable active metamaterials structures being developed today—winning investment from the likes of Khosla Ventures and Sequoia Capital, and commercial deployments in cellular access points and LG cellphones, she says.
“And we’re trying to do the same thing here,” Achour says.
Casse was for the last four years the area manager for metamaterials devices and applications within PARC, ultimately developing more than a dozen metamaterials technologies.
Metawave currently has a team of about 10 people, including part-time and contract employees, but Achour says she’s hiring rapidly, drawing from her deep network. “The [radio frequency] space is a very close community, and we all respect each other and value the innovation of each other,” she says, adding: “In the end, you attract them when you have a track record and you have credibility in this market.”