Building a better network for connected cars
In trials, connected-car technologies have proven effective in preventing accidents and easing the kinks in commuter traffic, most notably in the ongoing Connected Vehicle safety pilot program in Ann Arbor, Mich.
In the real world, however, connected-car networks face a huge obstacle – there aren’t enough cars that have the equipment to be connected. Fact is, the dedicated short-range communications (DSRC) devices that are being used for connected-vehicle trials aren’t yet being installed in new cars, nor are the roadside units DSRC requires being deployed outside of very limited trial areas.
“It will take 30 years before we see 100 percent of all vehicles having DSRC,” said James Martin, professor of computing at Clemson University.
Martin and his team have just received a two-year, $600,000 award from non-profit tech accelerator U.S. Ignite to develop an integrated network that could provide a bridge by enhancing DSRC with Wi-Fi and LTE cellular services, standards that are already being widely installed on new vehicles and, of course, cell phones.
Even when DSRC is ultimately installed in all vehicles, there are advantages to integrating other network standards, Martin said. “DSRC is an ad hoc wireless network, and if there is a string of 20 cars it might be 20 hops before the 20th car gets a warning.” That time lag could be critical, he said, and simultaneous alerts sent over integrated Wi-Fi or LTE could significantly reduce the delay.
In addition to developing middleware to integrate the network standards, the Clemson team will be developing applications to run on it. Initially, Martin said, the team will focus on producing two apps – one for incident detection and one to provide congestion warnings.
Anticipating that others will build applications to run on the network, the first question the team is grappling with is how much of the advanced networking capabilities to expose to the application developers.
According to Martin, the team is leaning toward using a software-defined network as the plumbing for application developers so they won’t have to directly deal with the underlying infrastructure.
“We know that developing and deploying vehicular applications is going to be very, very challenging,” he said. “It’s not a regular application. It is an application that is a distributed system, and there are pieces of the application that are running on different platforms with different capabilities.”
One of the biggest challenges for the project is the same facing connected-car technologies generally – getting enough vehicles equipped to make the system effective.
“What we’re trying to do is to effectively prototype a very small piece of what we anticipate the eventual deployment might look like,” said Martin. “So the challenge here is how do we get cars to participate?”
For starters, the Clemson researchers will be installing devices in South Carolina Department of Transportation vehicles that traverse the trial zone regularly. The team also plans to develop an app that would interact with the CANBUS microprocessor interface that is embedded in all cars manufactured since 2008. “And then of course we’d need to recruit volunteer drivers who understand the security and privacy implications,” Martin noted.
The team is planning to deploy hardware along 10 to 20 miles of South Carolina roadsides.
“Virtually every DOT in the country is at different stages of trying to figure out what to do with connected vehicle,” Martin said. “South Carolina is probably typical -- they have no clue. They’re using us as their first entry into trying to understand what connected vehicle is and its implications for their department.”
Posted by Patrick Marshall on Oct 20, 2015 at 1:26 PM