Ford Motor says it is skipping over semi-autonomous cars and going straight to fully self-driving ones, with no steering wheel or pedals, by 2021. Ford shares Google’s view, which is that it’s too difficult to safely hand over control from a semi-autonomous car to a human driver in an emergency situation. Instead, Ford says it’s putting all its efforts into fully autonomous cars and will have them ready in five years for ride-hailing and package delivery uses.
To achieve its 2021 target, the carmaker is investing in or collaborating with four start-up firms that have critical know-how. Artificial intelligence is a competitive field with many emerging players; here’s a little bit more about the companies Ford chose to work with.
Of all Ford’s new partners, Velodyne has been around the longest and is the best-known in automotive circles. It is the leader in light-detection and ranging (LiDAR) sensors, those giant spinning lasers on top of autonomous cars being tested by Google, Ford and others. LiDAR sends out a series of light pulses 360 degrees around the vehicle, gathering data that helps create high-definition 3D maps necessary for self-driving cars.
But Velodyne started out in 1983 in a very different business. Founder David Hall, a graduate of Case Western Reserve University, came up with a high-end, distortion-free subwoofer that set the benchmark for premium bass sound. The speakers were so popular with audiophiles that the company has been self-funded from the start, providing cash for Hall to pivot toward his next idea, LiDAR, in 2007.
Ford invested $75 million (as did China’s Baidu), which is the first time Velodyne has raised outside money, according to Chief Financial Officer Qing Lu. The $150 million cash influx will help Velodyne mass-produce more affordable LiDAR sensors. In just a few years, the cost of its LiDAR systems has fallen from about $70,000 to $8,000. Its newest system is about the size and shape of a hockey puck; it’s working on even smaller, solid-state LiDAR systems.
Ford discovered SAIPS while scouting for technology in Israel in 2013, and acquired the start-up more recently to strengthen its own artificial intelligence research. SAIPS’ founder and CEO, Udy Danino, worked for California-based Applied Materials before starting SAIPS in 2012. His company specializes in computer vision and “deep learning,” developing algorithms that process images, video and other signals. The software will help Ford vehicles learn and adapt to their environment while driving. Ford did not disclose what it paid to acquires SAIPS.
The scientist behind this year-old start-up is Sheila Nirenberg, a professor at Cornell University, who is developing a bionic eye to restore sight to the blind. Her breakthrough was finding a way to transmit a visual code directly to the brain, bypassing a damaged retina. An autonomous car is a lot like a blind person, says Nirenberg. It, too, needs help processing visual codes when there is no human to interpret them. Ford thinks the technology can help cars greatly improve their own vision systems, and process information the same way human drivers would. She has created two spin-off companies at Cornell. The first, Bionic Sight, hopes to begin FDA human trials soon. Meanwhile, Nirenberg Neuroscience, launched in February 2015, has intentionally been flying under the radar, she says. It has raised $3.5 million to date from investors, including some former Goldman Sachs executives, one of whom introduced her to Ford CEO Mark Fields. Ford now has an exclusive licensing agreement with the company.
Civil Maps was founded in 2014 by a group of computer scientists and robotics specialists from Stanford and the University of California-Berkeley to develop advanced 3D mapping for fully autonomous vehicles. Before working on self-driving cars, Chief Executive Sravan Pattagunta developed software for smart TVs. Civil Maps’ artificial intelligence software collects 3D data from LiDAR, camera and other vehicle sensors and organizes the information into machine-readable maps. Because the data doesn’t use as much bandwidth as traditional GPS maps, it is less expensive to transmit over existing cellular networks, making it easy to crowd-source, update and share road data in real time. Ford says Civil Maps’ innovative technique is scalable and more efficient than existing mapping processes. Ford was part of a group that invested $6.6 million in seed funding to accelerate the start-up’s technology. The Albany, Calif.-based company has raised about $9.6 million in total.