Spannende Einsichten, warum die Vehicle-2-Vehicle-Kommunikation beim autonomen Fahren, insbesondere bei schlechtem Wetter, so wichtig ist.
While humans have trouble driving in the snow and rain, these common weather phenomena present all sorts of disruptions to autonomous vehicles’ cameras and sensors. These limitations also help illustrate an important debate in the automation community over the importance of connectivity between cars and the amount of data stored within cars. At the moment, even a thin layer of snow troubles most (but not all) autonomous vehicles, since they can’t see the road. Further, since they use LIDAR — essentially radar using lasers instead of radio waves — autonomous vehicles’ sensor systems can be thrown off by lightly falling raindrops or snowflakes. What’s worse, cameras for detecting street signs, which are supposed to supplement radar and LIDAR, are easily rendered useless by fog. While skeptics may see these as reasons why autonomous vehicles are impractical, they reveal a deeper point about the impending wide-scale implementation of them. When we refer to a “smart” or autonomous vehicle, what do we mean? Do we mean making each vehicle so intelligent that it can solve any problem before it? Or do we mean that each vehicle is strongly connected with other vehicles and map systems, so that its intelligence is the result of the entire network? In the former camp, some are trying to ensure that autonomous vehicles can store most of their road data on board with them. Researchers are investigating minimizing the map data an autonomous vehicle needs access to in order to navigate the road. This would entail less detailed maps and would help cars avoid the costs of storing mass amounts of data, but it has its own downsides. Most pressingly, it can’t deal with novelty. With a relative lack of information, roads suffering with inclement weather conditions would be difficult for a car with such a system to effectively navigate.