When driverless vehicles become publicly available in just a few years, some people are wondering how they will respond to animals that enter onto roadways.
Swedish car company Volvo says that it plans to make its driverless cars adaptable to rural roads that are known for having many wild animals. In Sweden, drivers are often killed when their car strikes wayward moose.
Some people are even hopeful that driverless cars will be safer and more efficient when it comes to wild animals than error-prone humans.
According to officials from Volvo, this is achievable. Driverless cars will supposedly drive slower on roads known for containing frequent wildlife, and they will be able to identify animals and quickly respond to their presence.
It’s not just Sweden where roadkill is a problem. In the United States, hundreds of millions of animals die every year because of vehicular collisions. Many of these incidents involve large animals that can cause serious harm to the people within the vehicle. There’s also the financial cost, which was $4 billion last year.
While much success has been achieved in getting self-driving cars to safely interact around human-driven vehicles, working around animals has proven to be a more difficult task. Nature is unpredictable, and animals can often behave in erratic manners.
Roadkill expert Fraser Shilling said, “Even if we develop the perfect automated recognition and avoidance system, you still have an imperfect ecology and wildlife behavior system, so there’s maybe still too much chaos. If your subject, the animal, is sprinting out into the road in such a way that you can’t stop fast enough, does it really matter how perfect the car is?”
Driverless cars will certainly be able to recognize large animals. Anything that triggers the lasers will be treated as a hazard, and efforts will be taken to avoid it. Many of the lasers are so advanced that they can even identify what is in the way. Since animals contain flesh, they will be identified as a human pedestrian.
However, small animals, such as squirrels, might not be so lucky, as they are likely to miss the radars of the vehicles.
Still, what is difficult for developers is how to predict the movement of the animal. Humans can often observe the animal and account for speed and direction. The computers for self-driving vehicles cannot do this as easily.
Senior technical leader for Volvo Erik Coelingh said, “This is beyond state of the art. We can only make a rough prediction of the motion of the elk, based on its current position and speed. So when an animal is standing still we need to assume it will keep standing still until we see a movement.”
Indeed driverless vehicle researcher Andy Alden says that the situation is particularly complex. During a study that he conducted in conjunction with Toyota, his team was unable to successfully predict the movements to be taken by animals.
Alden said, “But there certainly are some things you could drop into an algorithm, like time of day, time of year, the kind of environment along the road, the width of the road, the amount of traffic on it. There are lots of parameters that affect your likelihood of encountering an animal on the road.”
Another idea involves putting sensors on the roads themselves in order to detect movement. When the sensors detect animals, they would alert nearby autonomous cars about such activity, allowing the cars to take action.
In the end, many experts believe that self-driving cars will eventually be better at avoiding animals than humans. It’s just a matter of getting there. Like almost every other aspect of this exciting new technology, more work still needs to be done.