AVs still need better tech — and more data — to drive in bad weather
A car in the Auvergne region of France in October 2018. Photo: Thierry Zoccolan/AFP/Getty Images
Most AVs use a combination of lidar, radar and cameras to survey their environment, but these technologies can fall short in even mildly inclement weather, and few AV systems meaningfully incorporate external data on weather and road conditions.
The big picture: Fully automated AVs will need to be able to pull and analyze information from various sensors on and off the vehicle to safely navigate. Despite industry optimism, however, getting this technology up to speed will likely take years.
Details: AVs can struggle to collect accurate data in adverse driving conditions, let alone make use of it.
- Ice, snow, water, and even dust or fog can obscure camera lenses.
- They can also cause a vehicle's lidar to interpret anything obscuring the sensor as an obstacle, triggering a sudden brake.
- Radar works proficiently in poor weather because its can penetrate thick fog, rain and snow. But it doesn’t offer enough detail for navigating busy roads.
Where it stands: Carmakers have begun to develop and test technology to address the challenges posed by weather.
- Google has fitted windscreen wipers onto vehicles lidar domes to improve detection accuracy, but in particularly adverse weather its cars have to pull over and wait for better conditions.
- Ford and the University of Michigan recently announced that they developed 3D map technology that augments lidar's capabilities when inclement weather interferes with sensors.
- Volvo is testing cars in Sweden to try to identify and address sensing challenges associated with low visibility, icy roads and heavy snow.
Connectivity will be key, including data transmitted by other vehicles.
- Road weather information systems, which measure weather and pavement conditions, also produce highly accurate data that should be made accessible to AVs.
What to watch: The American Meteorological Society and other organizations are stepping in to help facilitate discussion regarding weather data requirements, such as how often it should be transmitted and what level of precision is needed for AVs to operate safely.
Kevin Petty is the chief science officer at Vaisala and a former meteorologist at the National Transportation Safety Board.