Bad weather can be difficult to navigate. Heavy rain, snow, and fog can make it difficult for an audiovisual system to see the world around it. Mirrors can be created by wet roads. Sensor data can be damaged by fog. Ice can form on the lidar, making it hard for the sensor to send out laser points.
The company that wants to deploy fleets of robotaxis all over the country has quietly invested in weather research since it began as a "moonshot" project. The company has gotten better at detecting and predicting the weather in the cities in which it operates, and has even created a first-of-its-kind fog map for San Francisco.
Since its beginnings, the company has quietly invested in weather research.
The map is a result of millions of data points collected by Waymo's fleet of autonomously driving vehicles. Waymo is able to create a new meteorological "metric" that it feeds to its "Waymo Driver" in order to aid its decision making.
The company describes its vehicles as mobile weather stations, according to a member of the company's weather team. That is the way they are functioning.
The level of on-the-ground accuracy will become more important as Waymo gets closer to deployment of fully self-driving vehicles. After getting approval from the California Department of Motor Vehicles, the company is close to offering rides in San Francisco that are only for riders.
Our vehicles are described as mobile weather stations.
Arizona had a sunny, dry environment and was the location of the early testing. In the past few years, the company has expanded its testing to include more difficult conditions, such as snowy Michigan and foggy San Francisco. Arizona had its own edge cases, like giant dust clouds called "haboobs"
The company said it saw value in gathering more weather data. Real-time weather information can be found at weather stations. Aviation safety and climate monitoring applications are supported by more than one airport.
The progression of coastal fogs can be tracked by the fleet using this map. It can detect rain that is not visible to the National Weather Service's local radar. The company says that the weather observation capabilities allow them to see where the weather conditions are beginning to get worse or better. We will use the weather maps to create similar maps for additional cities as we scale.
We understand in real time the actual weather conditions that are impacting our vehicles in a hyper local context. That's something that hasn't been done before.
Robert Chen is the product lead for the weather team. It makes sense that the San Francisco Bay Area is important to the company.
Chen said that they are able to do light fog. There was some rain. By incorporating more hyperlocal, real-time data into its prediction and decision-making capabilities, the company expects that its vehicles will be even more successful in navigating through soupy conditions.
The goal is to build a vehicle that can reliably and safely drive through all types of bad weather in a bid to beat human drivers who are over confident in their own ability to see through thick fog.
This ability to understand the actual weather conditions has been very important to us.
All Rights Reserved © 2024
