Drivers may one day be able to find a parking space on their smartphones instead of on the street if a new system envisioned by University of Michigan researchers becomes reality. It would use radar sensors that are already built into many vehicles to create a crowd-sourced, real-time map of parking availability. The map could be linked to a smartphone app, enabling drivers to find street parking more easily or even reserve a spot in a lot or structure before they hit the road. The system could also enable autonomous vehicles to find parking on their own.
The research team, led by U-M Industrial Operations Engineering professor Romesh Saigal, recently conducted a preliminary test of the concept on U-M’s North Campus, circling a parking lot in a radar sensor-equipped test vehicle to capture data. They’ll then compare the radar sensor data with video footage of the test to determine just how well the car can spot empty parking spaces. If the research goes as planned, the team hopes to have a smartphone app ready for use within four to five years.
“Up to 30 percent of traffic volume in cities is due to drivers who are circling for a parking space,” Saigal said. “A system like this could help reduce congestion, reduce the environmental impact of driving and make urban life less stressful.”
Automakers are already building radar sensors into more and more cars. They’re used to sense nearby objects, generating data for features like automatic emergency braking, adaptive cruise control and automatic lane keeping. The team believes that those sensors may also be able to spot empty parking spaces as cars cruise streets and parking lots during normal driving. The combined data from thousands of drivers could be streamed to the cloud, creating a real-time database of parking availability. This, in turn, could be fed into a smartphone app that drivers could use to remotely grab a parking space.
For now, the radar sensors are mainly limited to high-end vehicles, but Saigal says they’ll inevitably trickle down to more and more vehicles, much as airbag and anti-lock braking technologies did in past decades.
“Today’s vehicles have more and more ways to collect data, and for industrial engineers like us, it’s a huge opportunity,” he said. “We’re asking ourselves what else can be done with that data, and this project opens some very intriguing new possibilities.”
U-M Industrial Operations Engineering graduate researcher Qi Luo says one of the key advantage of such a system could be low cost, since it relies mainly on software that would add functionality to existing hardware, rather than creating new components. He says this approach is far more cost-effective than other automated parking systems that are under development, which generally rely on costly hardware like in-pavement sensors.
“Using software to add capabilities to vehicles is a relatively inexpensive approach that we’re seeing more and more of in the auto industry,” he said. “And when you combine that new software with each car’s ability to collect data, all sorts of things become possible.”
One advantage of a software-based approach is the opportunity to integrate with other applications like calendars and restaurant apps. While details about any eventual system are still sketchy, Luo says a pilot version would likely be rolled out in Ann Arbor first, spreading to larger cities once it’s perfected.
“The great thing about relying on software is that it’s flexible and easy to adapt,” Luo said. “For example, it’s easy to imagine how a parking app could integrate with map and calendar apps to automatically find parking for every segment of your day. It also creates new ways for parking lot owners to sell parking and for cities to manage parking more efficiently.”
Luo envisions a system where drivers volunteer to share the data collected by the sensors in their vehicles and receive access to the smartphone app in return. Privacy concerns should be minimal since the data won’t be traceable to individual drivers. From a data collection perspective, he says the biggest hurdle may lie in convincing manufacturers to share the details of their proprietary systems.
“Each manufacturer has their own devices and their own data protocols. And for a system like this to be effective, it needs to work across those different systems,” he said. “Getting manufacturers to share how their systems work may be a challenge, but I think ultimately the benefits to drivers will convince them to cooperate.”
The team is conducting the project in conjunction with principal investigator Robert Hampshire, a research assistant professor at the U-M Transportation Research Institute.