Researchers are developing machine-to-machine (M2M) communication technology that allows cars to exchange data with each other, meaning vehicles will soon know what the cars all around you are doing on the highway.
Your car, for instance, could "see" the velocity of nearby vehicles and react when they turn or brake suddenly. And with computer algorithms and predictive models, your car will be able to predict where other vehicles are going and measure the other drivers' skills -- ensuring you're safe from their bad moves.
"We're even imagining in the future cars would be able to ask other cars, 'Hey, can I cut into your lane?' Then the other car would let you in," said Jennifer Healey, a research scientist with Intel.
Intel is working with National Taiwan University on M2M connectivity between vehicles as a way to make roads more predictable and safe.
"Car accidents are the leading cause of death in people 16 to 19 in the United States. And 75% of these accidents have nothing to do with drugs or alcohol," said Healey, who delivered a TED Talk on the subject in March.
She recounted her first accident when she was a young driver: The driver she was following on a highway slammed on his brakes and the resulting collision totaled her car. "I think we can transform the driving experience by letting our cars talk to each other," she said.
That idea came from caravanning, Healey said, citing an available, but-not-yet-deployed technology that uses direct line of site infrared (IR) and a range finder in order to automatically adjust the speed of cars so they can travel at a measured distance from each other. In other words, they're electronically tethered to one another.
Instead of using IR, the researchers wanted something that is omnidirectional. They tried radio communications, but quickly discovered that omnidirectional radio signals tend to bounce off vehicles, making them unreliable at high speeds.
So Healey and university researchers began using unique Internet Protocol addresses for vehicles, which would allow them to be instantly identifiable to nearby cars around on the same network.
"Imagine a group of cars traveling down the road together as an ad hoc network," she said. "Let's say you are three cars ahead of me and I get those IP packets that say I'm the packet from the blue car whose GPS position is here. Now I can associate my position with the unique ID of that physical blue object."
Along with a steady stream of data a bout the GPS location of cars around you, your car could also know drivers' intentions.
"I could [upload] my route to the cloud and, for example, let cars around me know I'll be on Rte. 101 for the next 10 minutes, and then I'm going to exit," Healey said. "You're augmenting on-road perception."
With a large enough cloud infrastructure, driver history could also be added, allowing cars to adjust their distance based on the safety record of other drivers. For example, a vehicle might identify a problem driver and simply monitor his or her car more carefully than other vehicles that have not been flagged.
"The car could passively let the driver know that red Jetta is someone you may want to watch more closely," she said.
Healey said the technology to create an automobile cloud network is readily available, but it's the reliability and scalability that remains unproven.
One obvious issue is bandwidth. Wireless communications vary by region, so while the system might work well in an urban setting, in a more suburban or rural area radio communications might be too slow to transmit accurate data.
Another problem is speed and traffic congestion.
"So if you're driving at 85mph, there is a physical problem of transmitting radio packets fast enough to exceed your speed such that other people can get it and react to it in time," she said. "So you'd have to start publishing a plan to go 85mph in my lane up Rte. 101. So I want to announce to cars 10 miles ahead of me that I'm doing that."
Of course, drivers may not want to publicize their plans to exceed the speed limit. "Law enforcement doesn't tend to like 85mph lane splitters," Healey said.
In additiion, the more vehicle there are, the more complicated the data exchange on an ad hoc network, Healey admitted.
"I can show you a Taiwan intersection with 100 cars coming into it. That's a problem," she said. "We're doing it for three cars, but can we do it for 100? [If] you can do this in a Taiwan intersection with four lanes and scooters coming across ... then you have a real situation."