Research finds driverless cars can’t easily predict bicycles’ unexpected traffic patterns or pick up on their paths, putting cyclists at risk for a crash
If you’re on a bicycle, will driverless cars be capable of detecting you — and avoid hitting you?
As it stands now, not exactly.
A future that puts driverless cars on America’s roads is getting closer, but it’s not there yet. Big problems still exist. Being able to safely detect bicyclists are among the most glaring.
Driverless cars still not in the right lane when it comes to bike safety
As an auto accident attorney who’s been a big proponent of driverless cars, I’m optimistic that driverless car technology will vastly decrease the number of people who are killed and seriously injured every year in car crashes and car/bike crashes. It won’t be perfect, but it will be better.
Bikes are big problem for driverless cars
“A self-driving car that doesn’t see 26% of the bicycles in its path isn’t going to cut it.”
The Forbes story quoted above explains that the technology behind driverless cars is still not where it needs to be when it comes to detecting and reacting to bicycles.
For example, self-driving car company Waymo (a Google spinoff) demonstrated its driverless car system at the 2017 North American International Auto Show in Detroit. Its system can detect and calculate short, medium and long distances to determine how fast the driverless car and other vehicles on the road — from fast-moving cars to slow-moving targets like bicycles — are approaching each other.
But, Forbes said, this system and others like Deep3DBox haven’t been able to make a reliable prediction for where the bicycle is going.
Citing the journal IEEE Spectrum, Forbes said that Deep3DBox — a network developed by researchers at George Mason University and Zoox, a company developing autonomous taxis — correctly identified 89% of cars on an industry-recognized benchmark test, but only identified 74% of bicycles.
Hence, the 26% not “cutting it.”
Working on a smart choice for driverless cars
We all know that a bicycle is much different than a car or truck in terms of size, but consider how and where it maneuvers.
If you’ve ridden a bike, you’ve most likely gone in jagged patterns along your way, making swerves when you turn. Some of your decisions can appear random or impulsive. As Forbes author Kevin Murnane explains:
“Bicycles are also fast, agile and prone to sudden and unexpected movements, which make the [driverless car’s] route-prediction problem even more difficult.”
There’s at least one bike line on the market that might be able to make a difference: the “smart” bike. NPR recently reported how engineers are coming up with getting these bikes to feed information to driverless cars through apps and built-in technology. As Anthony Rowe of Carnegie Mellon University explained:
“What we’re trying to do is put as much instrumentation on a bike as we can to see if we can predict how it’s going to move in the future, so that it could, for example, signal a collision warning system on a car.”
If such added technology also means added safety from, or with, driverless cars, frequent bicyclists may find investing in a smart bike — which cost an average of $1,000 each, with some ranging as high as $4,000 — to be worth it.
But the rest of us who have conventional bikes? Any one of us could be in that 26% missed by Deep3DBox.
Just as I wrote when Google had a self-driving car accident with a bus — where, thankfully, no one was injured — work still needs to be done before we can all trust driverless cars to drive us safely into the future.
Or, in bicyclists’ case, before we can trust them to drive alongside us.