Autonomous car testing in the dry, sunny climate of Arizona has led to some problems for companies that have been approved for testing in the state. Yet, a recent study in Michigan had results that point to even further problems at self-driving technology.
Some of the factors that the study indicated affected autonomous cars included rain, whether it was a light drizzle or a heavy downpour, as well as cold temperatures that lead to leaves falling from trees onto the street. Although the weather in Michigan is not replicated by many other places, it does pose some questions for the ability of self-driving cars to adapt to different climate patterns.
Michigan State University feels a connection to this growing industry because of its proximity to Detroit and the city’s history in relation to car manufacturing. The university has started the Connected and Autonomous Networked Vehicles for Active Safety (CANVAS) Program to study how autonomous cars handle the tasks of not hitting things they shouldn’t. Part of this studies the effects of the weather’s impact on cameras, sensor rays, and other technology that guide self-driving cars.
While the final results of the study have not yet been released, they did make it clear that autonomous cars have a long way to go before they can become the norm.
The worst of the impeding factors to this technology is rain, especially when it comes to pedestrian and cyclist safety.
“When we run these algorithms, we see very noticeable, tangible degradation in detection,” Hayder Radha, the MSU professor of electrical and computer engineering who oversaw the project said. “Even low-intensity rain can really create some serious problems, and as you increase the intensity, the performance of what we consider state-of-the-art mechanisms can almost become paralyzed.”
This is not the first time problems have appeared with automotive features designed to aid drivers. Lane-keeping features also faced problems during initial testing due to rain.
High-resolution mapping used in navigation also loses efficiency with the changing of seasons. As foliage falls from trees, the landscape doesn’t look the same leading to further problems.
“You can imagine in environments where there are a lot of leaves on trees or on shrubs close to the road, they are an essential part of the map,” Radha said. “So summer and winter are completely different. When they fall down in winter, you have nothing to work with.”
Furthermore, cold temperatures, especially ones under ten degrees Fahrenheit, led to an uptick in “noise” registered by lidar sensors. Trying to compensate for this problem could lead to a dumbed-down system which is even less safe. Radha also said that he is worried about the potential hazards that snow on the roads would pose to driverless cars and is curious to see how manufacturers plan to work around this in the future.