SAN FRANCISCO: Waymo has long sought to understand how weather can affect the performance of its autonomous vehicles.
Just as a human can quickly reduce speed or adjust braking depending on whether it's raining or not, or whether visibility is good or bad, Waymo's driving system must also be able to react to the slightest change in weather conditions.
Unfortunately, the weather can alter the car's perception of the environment. Wet roads can create reflections on the cameras, for example, while fog, mist and rain can slightly affect the data readings of the different sensors.
Waymo is now able to measure the presence (or absence) of raindrops on the windows to assess different weather conditions.
By combining data from the hardware sensing suite with data from weather visibility sensors, the vehicle can analyse the weather around the vehicle in real-time to establish whether it’s foggy or raining, for example, and to what intensity. Waymo's automated driver can then adapt to the identified conditions.
These mobile weather stations also provide an accurate view of the climate evolution over the whole of San Francisco.
Over time, Waymo has been able to develop an accurate fog map of the city, showing how coastal fogs that arrive from the Pacific Ocean progress before dissipating in the morning.
In future, the idea is to create similar weather maps for Phoenix and other cities that Waymo will enter.
