Elon Musk famously stated any firm counting on lidar is “doomed.” Tesla as an alternative believes automated driving capabilities are constructed on visible recognition and is even working to take away the radar. China’s Xpeng begs to vary.
Founded in 2014, Xpeng is considered one of China’s most celebrated electrical automobile startups and went public when it was simply six years outdated. Like Tesla, Xpeng sees automation as an integral a part of its technique; in contrast to the American big, Xpeng makes use of a mix of radar, cameras, high-precision maps powered by Alibaba, localization methods developed in-house, and most lately, lidar to detect and predict street situations.
“Lidar will provide the 3D drivable space and precise depth estimation to small moving obstacles even like kids and pets, and obviously, other pedestrians and the motorbikes which are a nightmare for anybody who’s working on driving,” Xinzhou Wu, who oversees Xpeng’s autonomous driving R&D middle, stated in an interview with TechCrunch.
“On top of that, we have the usual radar which gives you location and speed. Then you have the camera which has very rich, basic semantic information.”
Xpeng is including lidar to its mass-produced EV mannequin P5, which is able to start delivering within the second half of this yr. The automobile, a household sedan, will later be capable to drive from level A to B primarily based on a navigation route set by the motive force on highways and sure city roads in China which can be lined by Alibaba’s maps. An older mannequin with out lidar already allows assisted driving on highways.
The system, referred to as Navigation Guided Pilot, is benchmarked in opposition to Tesla’s Navigate On Autopilot, stated Wu. It can, for instance, robotically change lanes, enter or exit ramps, overtake different automobiles, and maneuver one other automobile’s sudden cut-in, a standard sight in China’s complicated street situations.
“The city is super hard compared to the highway but with lidar and precise perception capability, we will have essentially three layers of redundancy for sensing,” stated Wu.
By definition, NGP is a complicated driver-assistance system (ADAS) as drivers nonetheless have to hold their palms on the wheel and take management at any time (Chinese legal guidelines don’t permit drivers to be hands-off on the street). The carmaker’s ambition is to take away the motive force, that’s, attain Level four autonomy two to 4 years from now, however real-life implementation will hinge on rules, stated Wu.
“But I’m not worried about that too much. I understand the Chinese government is actually the most flexible in terms of technology regulation.”
The lidar camp
Musk’s disdain for lidar stems from the excessive prices of the distant sensing methodology that makes use of lasers. In the early days, a lidar unit spinning on prime of a robotaxi may price as a lot as $100,000, stated Wu.
“Right now, [the cost] is at least two orders low,” stated Wu. After 13 years with Qualcomm within the U.S., Wu joined Xpeng in late 2018 to work on automating the corporate’s electrical vehicles. He at present leads a core autonomous driving R&D workforce of 500 employees and stated the drive will double in headcount by the tip of this yr.
“Our next vehicle is targeting the economy class. I would say it’s mid-range in terms of price,” he stated, referring to the agency’s new lidar-powered sedan.
The lidar sensors powering Xpeng come from Livox, a agency touting extra reasonably priced lidar and an affiliate of DJI, the Shenzhen-based drone big. Xpeng’s headquarters is within the adjoining metropolis of Guangzhou about 1.5 hours’ drive away.
Xpeng isn’t the one one embracing lidar. Nio, a Chinese rival to Xpeng focusing on a extra premium market, unveiled a lidar-powered automobile in January however the mannequin received’t begin manufacturing till 2022. Arcfox, a brand new EV model of Chinese state-owned carmaker BAIC, lately stated it might be launching an electrical automobile outfitted with Huawei’s lidar.
Musk lately hinted that Tesla might take away radar from manufacturing outright because it inches nearer to pure imaginative and prescient primarily based on digicam and machine studying….
