Self-driving taxi operator Waymo might have a way of getting around its vehicles’ problems operating in foggy San Francisco: by turning each into a roving weather station that can adapt to the conditions.
From a hardware perspective, it appears to be a simple trick, Waymo said in a blog post. Its current hardware – cameras, radar, lidar – are simply pointed back at itself.
The hardware and Waymo Driver’s AI engine “uses the raindrops on its windows – or lack thereof – to classify various weather conditions.” According to Waymo, the approach is similar to the way laser-based instruments have previously been used to measure fog, clouds, dust storms, and other weather phenomena.
Waymo said that by adapting the laser concept and pairing that data with “high-quality ground-truth data” from visibility sensors, it has designed “a quantitative metric about meteorological visibility.”
Armed with that metric, Waymo Driver can analyze the weather around its vehicle in real time. “Simply put, each Waymo vehicle operates as an autonomous ‘mobile weather station’, providing an unprecedented understanding of weather in the areas we drive,” Waymo said.
Ground-level San Fran weather: ‘Unpredictable’ is generous
When Waymo’s driverless taxis started operating in San Francisco last year, it was with a firm restriction: weather worse than light rain or light fog stopped service. In San Francisco, that can mean a Waymo taxi taken out may not be able to get you home.
“Even over short distances – such as between San Francisco International Airport and the Sunset district – local conditions can vary significantly,” Waymo said, adding that modern weather tools lack the specificity and precision required to operate in places like the Golden City.
It’s for that reason that Waymo Driver’s weather data isn’t restricted to one vehicle – it’s sent back to Waymo HQ to be used as a “first of its kind” city-wide fog map.
Waymo said the map allows its fleet to track the progression of fog as it rolls in from the Pacific and burns off as the sun rises, and can even “detect drizzle and light rains that lead to wet roads in situations that are invisible to the National Weather Service’s local Doppler weather radar.”
Now can we have that all-weather permit, please?
Along with using Waymo sensors to map out ground-level conditions, the company said it plans to use the data its vehicles collect to train the Waymo Driver in simulated environments so it can improve the AI’s performance “regardless of the time of year.”
Waymo said that the weather-sensing technology is being used now in San Francisco and Phoenix “and we’ll create similar weather maps for additional cities as we scale.”
While expansion is obviously on the books given that the company has announced plans to enter Los Angeles, it’s San Franciscans who are likely to be thrilled by the news. Residents who want to ride in a Waymo driverless taxi have to be part of Waymo’s Trusted Tester program, meaning most locals can’t hail a driverless Waymo cab.
If Waymo is able to improve its ability to operate in the unpredictable and often-foggy San Francisco weather, it might finally prove the practicality of operating robotaxis in places other than deserts. ®
- ai art
- ai art generator
- ai robot
- artificial intelligence
- artificial intelligence certification
- artificial intelligence in banking
- artificial intelligence robot
- artificial intelligence robots
- artificial intelligence software
- blockchain conference ai
- conversational artificial intelligence
- crypto conference ai
- deep learning
- google ai
- machine learning
- plato ai
- Plato Data Intelligence
- Plato Game
- scale ai
- The Register