UD researchers look at creating “smart cities” for driverless cars
Driverless vehicles may be the wave of the future, but what if? What if a system somehow built into our roadway infrastructure could tell those vehicles whether to speed up or slow down as they merge onto Interstate 95, or where to turn off Kirkwood Highway to avoid a traffic jam a half-mile ahead?
An experiment now under way in the University of Delaware’s Mechanical Engineering department is trying to figure out whether such a system might work and, if so, how it might be put together.
It’s serious business, but the work that’s going on in Room 222 of Spencer Lab looks like a lot of fun.
In a 20-by-20-foot floor space, Associate Professor Andreas Malikopoulos and a diverse 30-member team composed of two post-doctoral research associates, 25 graduate and undergraduate UD students and three high schoolers, has created a Scale Smart City, a carefully laid out model that bears at least a passing resemblance to a community like Newark, right down to the large UD logo imprinted in the center of a traffic roundabout. Some 35 battery-powered, computer-guided model cars and trucks equipped with tiny cameras traverse the mini city’s roadways, speeding up or slowing down as they try to keep a safe distance behind the vehicles ahead, stopping at intersections, and taking care to turn left or right only when the path is clear.
“We’re replicating city driving, about 30 miles per hour, using algorithms and trying to even out traffic flows,” Malikopoulos says. Unlike testing by Tesla, Uber and others that are focusing on producing safe autonomous vehicles, this project is seeking a means to control an entire community of vehicles, using computers and sensors to enable vehicles to “communicate with each other so they can avoid conflict,” he says.
“We’re building a super-flexible testing platform,” says Ben Remer, a graduate student who worked on engineering vehicle interiors for Ford before coming to the university last September.
“It’s an eye-catching project … and it’s kind of cool to jump into the future,” Remer says.
More importantly, he’s enthused by the potential magnitude of the project. “Working on an autonomous vehicle improves safety only,” he says. Studying the interconnected behavior of multiple vehicles raises the stakes – opening new paths to increasing energy efficiency while reducing both pollution and transportation times.
The testing must begin at the micro level – hence, the Smart Cities project. But the next step, advancing to the use of real vehicles in a controlled environment, perhaps on the university’s developing STAR Campus, might be only a year or so away, Malikopoulos says.
Those details remain to be worked out, but the university is already in talks with the Delaware Department of Transportation on how to move the project forward, he says.
Watching the model cars maneuver on the scaled-down roadway gives a sense of what is possible, but team members know it will take much more to make it a reality.
Before putting a system like this on the road, “industry standards need to be developed,” Remer says. And, after that, some sort of infrastructure would have to be installed, and that might well require an extensive network of computerized sensors and monitors plus some form of electronic communicator that could send messages to all the vehicles linked to the network.
“We’re at the bleeding edge. We’re not at the cutting edge yet. We’re still trying to understand the lay of the land,” says Dimitris Assanis, one of the two postdoctoral students working on the project. “Looking 20 or 30 years down the road, can we eliminate traffic light and stop signs? The beauty of this research [in a university setting] is that we don’t have to solve the problem in two or three years because we have to answer to our [corporate] investors.”
As the son of the university’s president, touting the benefits of campus research comes easily for Dimitris Assanis. And this subject is close to his family’s heart as well, as UD President Dennis Assanis, in one of his former roles, was director of automotive laboratories at the University of Michigan, which happens to be where Malikopoulos earned his master’s and doctoral degrees.
In his undergraduate and graduate studies at the University of Michigan, Dimitris Assanis focused on power and propulsion issues, studying ways of decreasing energy consumption and reducing harmful emissions in motor vehicles. In the Smart Cities project, he is learning about another tool to help achieve those objectives. “Instead of improving the efficiency of an engine, or of a hydrogen fuel cell, we can change the behavior of the driver and create additional efficiency outside the actual vehicle.”
While Assanis’s background naturally leads him to focus on fuel consumption and emissions, others involved in the project are just as happy to talk about how development of the Smart Cities concept can help reduce commuting times – not only in Delaware but around the world.
UD junior Yiming Wan, who spent his early years in a near Shanghai, China, with a population of more than 24 million, laments the traffic jams his mother encounters when she drives to work. “In any city in China, traffic jams can occur any time of day. For my mom, it can take an hour and 30 minutes,” he says.
Behdad Chalaki, a first-year doctoral student who did his undergraduate work at the University of Tehran, says traffic congestion in Iran’s largest city, population 8.8 million, can be just as bad as Wan says it is in China.
“You can easily get caught in traffic for an hour and a half,” he says.
“But, if you don’t get stuck in traffic, you will have a better day. If you start your day in traffic, you get frustrated,” he says.
UD junior Becky Williams, a graduate of Mount Pleasant High School in suburban Wilmington, acknowledges that development of Smart Cities’ concepts might help conquer her aversion to driving in congested urban areas. And she believes that the development of autonomous vehicles and computer-assisted infrastructure might relieve some of the tedium and reduce some of the dangers of traveling long distances at high speeds on interstate highways.
Working on the project has required all the participants to broaden their skillsets. Williams, for example, said she had previously done a lot of work with composite materials, but here she has learned more about design, electronics, robotic and controls.
Day in and day out, the team tests new approaches.
“We bought circuit boards off the shelf. We merged a couple of different circuit boards, made a custom chassis,” Assanis says.
Wan was given the responsibility of soldering the circuit boards together, one to run the car, the other to receive instructions from the computer, and to hook them up to a tiny battery and voltage regulator. “I’m the hardware guy,” he says. “If something goes wrong [with a car], they need me.”
Creating new algorithms, sending new sets of instructions to the vehicles, watching how they respond, it’s all part of the project. “We track energy usage, the level of the batteries, power levels, the time it take to go through the intersections” and much more, Assanis says.
Many questions remain to be resolved, not the least of which are identifying the most effective infrastructure system and determining the best means for transmitting information to vehicles operating within the system. Underground sensors and overhead cameras are two possibilities for the infrastructure. Another possible component, not yet tested, is a drone (or drones) that could hover over the Smart City, with its omnidirectional camera capability sensing bottlenecks and transmitting signals to vehicles on the road.
The search for answers can be challenging, but hardly tiring.
“There are so many variables,” Williams says.
“If we can improve the technology, the sensors, we can do all this stuff in the real world,” Chalaki says. “This is the future.”