Metawave was founded last month to create digital eyes that can discriminate objects and map the surroundings of autonomous cars. The company, which recently raised $7 million in its first funding round, plans to do that with a new type of radar.
The company says that its synthetic aperture radar will be capable of electronically steering signals in narrow beams, using control software embedded in a circuit board. This is possible because the radar's antenna is fabricated out of metamaterials, which exploit tiny repeating structures to bend radio waves in unnatural ways.
Metawave is also trying to improve the intelligence of radar, which normally only identifies location and speed. The radar will apply deep learning algorithms to infer the size and shape of objects, enabling it to discriminate a pedestrian, for example, from another car on the road. This way, the radar sends cleaner data to central processors, which can make quicker decisions about avoiding obstacles.
Metawave was started in August by chief executive Maha Achour, an electrical engineer that previously formed Rayspan, which went out of business trying to sell metamaterial antennas for smartphones. The other founder is chief technology officer Bernard Casse, a former manager at the Palo Alto Research Center (PARC), which Metawave spun out of.
In the future, autonomous cars will employ a wide array of sensors – including cameras, radar, and lidar – to provide redundancy. Cameras struggle at night and in rainy weather, while lidar have limited range and can interfere with other lidar on the road. Radar can compensate for these shortcomings, while other sensors compensate for the gaps in radar, which normally have low resolution.
“The auto industry has gone through cycles to find the best camera and the best lidar, and now it’s the radar era – time to find the best and smartest radar platform,” Casse said in a statement about Metawave's funding. The company is now looking to hire engineers. It has not said when it could release a final product.
Metawave is not alone in trying to give radar both tracking and mapping capabilities. Echodyne, which has raised $44 million from investors including Microsoft founder Bill Gates, is applying its metamaterials expertise to create short-range automotive radar. The company has not said when it will be released.
Oculii has built radar sensors that position other cars on the road in three dimensions, while Oryx Vision has raised $67 million for its coherent optical radar, which uses lasers to illuminate the road and measures the reflected signals to locate objects. The sensor resembles radar almost more than lidar because it treats the reflected light as a wave, not a particle.
Among the investors in Metawave’s fundraising were Khosla Ventures, Motus Ventures, and Thyra Global Management. Further out, the firm plans to tune the same metamaterials and algorithms for wireless antennas, which could connect cars to 5G networks.