WeRide now has over 100 self-driving cars in test operation. (Credit: WeRide)
WeRide, a Chinese robotaxi startup, recently compared its results testing self-driving cars in both Silicon Valley and Guangzhou, China. The report trumpets, “Test Mileage in Guangzhou 30X More Efficient Than in Silicon Valley.”
The efficiency gains are driven by the non-motorized traffic density of Guangzhou’s street compared to Silicon Valley’s, alongside more traffic violations.
WeRide’s “Road Test Data Comparison 2019 US vs. China” graphical summary
WeRide finds, “The number of cars per mile in Guangzhou and Silicon Valley is similar, but Guangzhou has 4-5x the number of pedestrians and bikes.” Taken together with WeRide’s measurement that “cyclists riding against traffic are 60X more common in Guangzhou”, much of the efficiency gain from testing in Guangzhou seems to come from handling rogue cyclists.
WeRide emphasize its safety focus, an important consideration when encountering the number of pedestrians and cyclists in Guangzhou. According to the report, “WeRide always puts safety first in its testing and operation. By accumulating more and better data in China for model training, WeRide can quickly iterate more stable autonomous driving systems.”
The comparison between Guangzhou and Silicon Valley is pertinent to other self-driving operations, which have to consider where to test. Many self-driving car companies, including Waymo, have focused their operations on relatively favorable geofenced locations, such as Phoenix, Las Vegas, and Silicon Valley. In these areas, a combination of sunny weather, wide streets, and good infrastructure helps the programs progress.
Other self-driving car operations have gravitated toward more challenging environments, such as San Francisco, Boston, Pittsburgh, and several Chinese cities.
In 2017, Cruise co-founder and CTO Kyle Vogt shared similar conclusions to WeRide, stating, “Based on our experience, every minute of testing in San Francisco is about as valuable as an hour of testing in the suburbs.”
The risk, of course, is that the environment will prove too challenging for early-stage autonomous vehicles. In 2016, an Uber
Favorable locations are hardly a guarantee for successful self-driving car testing, however. Tragically, the first Level 4 self-driving car fatality took place in 2018 on a wide, empty street in Arizona.
Around the world, self-driving efforts are tackling many different environments. Time will tell if there is a “right” place to start.
