Acknowledging that foul weather is a challenge for automated-vehicle (AV) operation, the CEO and founder of a company that creates highly accurate weather forecasts for industries around the world has his sights set delivering “road-level” weather forecasting that he believes will help AVs operate more often and more safely in poor weather.

Mark Flolid, CEO and founder of Global Weather Corp., said at October’s SAE Innovations in Mobility (IIM) conference that enhanced data will allow prediction of road-level weather conditions. This will permit AVs to operate for expanded periods in areas that experience weather conditions that can befuddle some or all of the sensors on which AVs rely for high-level automated functionality.

Flolid said having access to highly-accurate weather data is no longer good enough. “There’s a difference between ‘weather’ and weather at road level,” he said during a presentation at the IIM conference. Although his company already uses sophisticated modeling for its current weather data that it claims is the world’s most accurate, Flolid is anxious to lay the foundation for gathering much more granular data to enhance the models currently used to forecast road-surface conditions.

Road-level weather data, he said, could be integrated with high-definition maps used by AVs to not only make high-level autonomy safer, but broaden AV availability of at times of the year when foul weather most impacts AV functionality (see chart in gallery). In some cases, Flolid asserted, availability of high-level automated driving could be improved by as much as 90% if road-surface information was available.

Where would that data come from? From the vehicles themselves. Flolid hopes to interest automakers or other large-fleet operators in a strategy to share information from sensors already on vehicles. His ideal would be traction information. That sounds a lot like the vehicle-to-vehicle (V2V) data-transfer the industry has said it is pursuing, Flolid concedes, but he said V2V efforts haven’t typically eyed the exchange of mundane data such as vehicle traction- or stability-control engagements. “That info isn’t readily available,” he said.

An important factor is scale, he added. Random and sporadic information isn’t sufficiently robust to build a reliable picture of road-surface conditions at any moment. A major player – better yet, many major players – are vital to deliver the sheer number of vehicles required. He said several German automakers have expressed interest, as has at least one U.S.-based automaker 

The data Flolid wants to leverage wouldn’t be useful only for AV operation. He said it also would be invaluable in verifying the weather-prediction models already in use, particularly for edge cases and for rapidly changing conditions.