Will Autonomous Vehicles Keep Cyclists Safe?

For years, there’s been a lot of talk about the potential of self-driving cars. There are many arguments for and against removing the human element from behind the wheel, but that doesn’t remove human error and interaction outside the vehicle. As cyclists and pedestrians, how dangerous are the roads with driverless cars, and how far away is the day we see them on the streets?

We’ve spent a lot of time over the past few weeks looking at the epidemic of pedestrian fatalities. Drivers aren’t that much safer, and over the past two decades, many experts have pointed to the evolution of self-driving cars as the answer. By removing distracted, tired, drunk, or inexperienced drivers from the equation, we can remove the uncontrollable human elements that lead to crashes. Artificial intelligence has made massive strides in the past five years, while connectivity issues and ‘lidar’ have made the proverbial eyes of these vehicles nearly as safe as human drivers in some circumstances. 

Unfortunately, those successful scenarios don’t reflect real-world driving conditions. These vehicles have been programmed and designed to handle things like other cars, signs, stop lights, roundabouts, and other road infrastructure rather well. But what they don’t deal with safely enough are often the most vulnerable road users like pedestrians and cyclists. 

In March of 2018, an Uber test vehicle killed a woman walking her bicycles across the street. The car simply didn’t recognize her. Failing to register as a pedestrian, it didn’t hit the brakes until milliseconds before impact, which was far too late. Her death sparked outrage and put an immediate halt on many company’s previously very public testing of similar systems. 

Google, Uber, and a number of other companies have permission to test many of their vehicles on public streets, and while they’re doing so with a much lower profile than in years past, they’re still forcing prototype technology to make real-world decisions with real-world consequences. Google’s autonomous division, Waymo, registered over 10 million miles in a single month. Lyft said that it expected to have all of its vehicles be completely self-driving by 2022. The reality isn’t that these vehicles will be on the roads soon, it’s that they are right now, and in ever-increasing numbers. 

Critical to the arguments for this real-world testing is that it takes time to learn, according to designers. The artificial intelligence that powers these vehicles relies on data to form its driving habits. It needs to ‘see’ all of these scenarios not just once, but hundreds and maybe even thousands of times so that it can learn to predict, anticipate, and react to road users like cyclists and pedestrians. Advocates would say that the only way to achieve the vision of autonomous vehicles becoming safer than human drivers is by letting them go, learn, and shape the roads. 

To that end, companies like Waymo and Toyota are building their own small cities and urban landscapes to allow for testing without putting the public at risk. Uber has taken a big step backward in it autonomous technologies, restricting testing to a single location, Pittsburgh, and only using AVs during daylight hours. 

With road fatalities increasing slightly, AVs do offer hope. They could be a vital tool in reducing those deaths. The Department of Transportation attributes up to 94% of car crashes to human error, and these autonomous vehicles could greatly reduce those instances. However, we are still years away from completely autonomous vehicles. Today, many cars on the road use limited driving assistance to alert the driver of dangers, and there may be some applications of semi-autonomous cars on the road that have a driver ready to take the wheel. Fully autonomous still needs more development and the correct testing and regulation from governments. 

For now, we rely on information, education, road design, and human compassion to stay safe as cyclists and pedestrians. As long as people choose non-motorized transportation and recreation, we need to value the human element on the streets and prioritize people over vehicles.