Petros Ioannou had lost the battle with Santa Monica. He and his wife retreated.
“It was the worst experience I had with parking in Los Angeles,” Ioannou reflected. “We didn’t know there was a farmer’s market event happening. Roads were closed. We got lost — we didn’t know where the parking structures were. We’d always parked in a particular spot and now the configuration had changed.”
“We wandered around, got frustrated and went away.”
Now, due to a recent USC Viterbi partnership with Audi and its Electronics Research Laboratory, those sorts of memories may be a thing of the past, a quaint reminder of life before smart cars and “intelligent assist.”
This fall semester, Ioannou — a professor in the Ming Hsieh Department of Electrical Engineering — and his students in the Center for Advanced Transportation Technologies (CATT), received some new wheels — a black Audi A8 fully equipped with a broad sensor suite that includes radar, lidar (for light detection and ranging), cameras and WiFi. For the next two years — the Audi smart car will be the USC test bed for everything from parking assist, adaptive driver assistance systems and personalized navigation.
INTELLIGENCE — AUDI-STYLE: Professor Petros Ioannou and his research students have taken the lead on a four university research partnership with Audi's Electronics Research Laboratory. The new car, delivered earlier this fall, will be the proving ground of the latest ideas in parking assistance, adaptive cruise control and personalized navigation.
Along with USC, the University of California, Berkeley, the University of California, San Diego and the University of Michigan are also contributing to the three-year project.
With nearly 10 million souls in Los Angeles County alone — the CATT team’s research has enormous implications for managing traffic flow, driver comfort, mitigating accidents and, certainly, finding parking at Staples Center on a Friday night.
Imagine a different Santa Monica driving experience — one where the weekend farmer’s market is not the death knell of an afternoon. A scenario where the driver can input a destination, and the car will run algorithms to decide — based on the time of day, the driver’s cost preferences, scheduled events, the number of available lots — where to park. Indeed, as currently envisioned, the car might even reserve your spot and pay remotely before you’ve even left the driveway.
Personalized driving is another big driver of the research. For all the smart car’s future intelligence, it will remain more of a 300 horsepower personal assistant that caters to the personalities of multiple drivers.
Fully equipped with radar, lidar (light detection and ranging), cameras and WiFi, the Audi smart car will be the focus of the next two years: easing the congestion, dangers and uncertainty of drivers in an urban megacity.
“The car would learn my driving style — acceleration profile, lane usage and level of intervention — and tune its assistance systems to my style,” Ioannou said. “Then, if you switch to my daughter, who has much shorter reaction times and a much faster driving response — typical of a young person who thinks a car is like a toy — the assistance she receives will be different than mine.”
What does that mean? It means that the car — whether it’s approaching a stop sign, trailing a car, or veering dangerously into another lane — knows your reaction times better than you do and will customize a response accordingly.
In addition, the CATT team’s research will go beyond the limited traffic data of say, an iPhone’s Sigalert app, to collect data from all possible sources: life data, historical data, unpredictable events, even simulations that can do fast forward events.
Said Ioannou, “Let’s say there’s an accident on a particular link of the highway — we could try to predict through simulations how traffic will look 10 minutes from now. We know the traffic patterns and can give better guidance to the driver, how best to go from point A to point B.”
So, for example, when the unpredictable does happen — last week’s tanker truck explosion on the 60 Freeway in Montebello — the smart car’s simulator will imagine a virtual traffic network, based on real time and historical data to forecast the future. When a massive river of steel is diverted, where does it go?
For the students, the next stage is January, where they’ll deliver the first real demonstration of intelligent parking and driver’s diagnostics. To do this, however, means solving some big challenges: integrating Audi’s existing software tools into the navigation, getting different software and databases to talk to one another, receiving the most accurate traffic data and, of course, understanding the various ways of estimating driver and vehicle behavior.
That last part means taking the new Audi for a spin — in the rain, in the dark, in the daytime — collecting reams of data. Ioannou’s four graduate and Ph.D. students have driven the A8 in the University Park neighborhood, all over Los Angeles and even up to San Francisco.
For Ioannou, a native Cypriot whose first car was a Matador — made by the now defunct American Motors — flash has never been paramount. He loves the dynamics of a motor vehicle. His students’ work will likely not be seen in the color and the shape of the new smart cars, but in the underlying algorithms, the unseen presence that solves your parking, predicts the future or saves you from the head-on collision.
Lacking that, in a city like Los Angeles, it may also be your best defense against the scourge of farmer’s markets.
