Argo AI just spent Ford's cash on a key driverless sensor tech
Autonomous car developer Argo AI has snapped up a key developer of the sensors driverless vehicles rely on to see the road, as the industry ponders on how to make self-driving cars affordable. Argo AI was the brainchild of former Google and Uber experts, and caught the attention of automaker heavyweight Ford in the process. Earlier this year, Ford said it would invest $1bn in the startup over the next five years.
In doing so it became Argo AI's majority shareholder. Ford has already committed to getting a commercial self-driving vehicle on the road in 2021, part of a push to embrace new technologies that will also see a hybridized version of the F-150 truck, which for decades has been America's best-selling vehicle.
Now, Argo AI is splashing some of that Ford investment around. It has acquired Princeton Lightwave, a company specializing in the development and commercialization of LIDAR sensors. LIDAR is the laser rangefinder technology which allows autonomous cars to build up detailed 3D point-clouds of their surroundings by bouncing light off them.
Most driverless car projects count LIDAR among their sensor suite, generally combining it with camera, radar, ultrasonic, and other technologies. The problems with the laser-based system, however, has always been its size and cost. Indeed, only a few years ago autonomous car developers complained that a single LIDAR array was the single most expensive part of a driverless vehicle.
Since then, bringing down the price has been a key area of interest for LIDAR specialists like Velodyne and others. After all, while a handful of prototype vehicles for research and development purposes can be expensive, if the mass-market is to adopt driverless cars it'll need to be competitive on dealership forecourts.
Key to Princeton Lightwave's value is so-called Geiger-mode LIDAR, a system to which it owns the intellectual property. That relies on single-photon detectors operating at near-infrared/shortwave-infrared wavelengths, and which the company boasts can see further out than any competing LIDAR sensor. Its latest GeigerCruizer sensor, for instance, promises the ability to track 60 mph highway traffic at 350 meters, potentially more than three times the distance of rival hardware.
According to Argo AI CEO Bryan Salesky, formerly of the Google autonomous car project that went on to become Waymo, that's one of the reasons why this talent-add makes a lot of sense. The company has already been working on its own sensor hardware, and the Princeton Lightwave technology will improve "the interface between sensor and software," he said today.
"Princeton Lightwave's technology will help us unlock new capabilities that will aid our virtual driver system in handling object detection in challenging scenarios, such as poor weather conditions, and safely operating at high speeds in dynamic environments," Salesky explains. However, while Argo AI will use it with its own sensors, the tech could also improve the performance of third-party LIDAR hardware.
That keeps the door open to Argo AI not only developing self-driving technologies that Ford can use, but to making them a new source of licensing income. From the start, the automaker has been clear that it expects part of Argo AI's business to be providing other companies – in the auto space and further afield – with the tech to offset human involvement.
"In the future," Ford said back in February 2017 when it announced its investment, "Argo AI could license its technology to other companies and sectors looking for autonomous capability." That could include factory robots, airport hardware, and warehouse equipment.
"We are constantly exploring how to increase the range, resolution and field of view of LiDAR," Argo AI's Salesky says, "but we're also looking to lower costs and manufacture these sensors at scale."