Searching for a parking spot can be one of the most aggravating parts of driving, but now there may be a way for your car to do it for you.
BMW will demonstrate something called Dynamic Parking Prediction at the TU-Automotive Detroit connected-car conference, using an i3 electric car.
The prototype Dynamic Parking Prediction system builds on BMW’s existing Connected Navigation, which uses real-time traffic information to predict delays and help to route drivers around congestion. The carmaker believes the same tech can also find parking spaces.
The system uses movement data from vehicle fleets to predict when spaces will be open. An algorithm then compares these movement patterns to a digital map of all of the available public parking spaces within a given city, and tries to figure out which ones are most likely to be up for grabs.
This setup relies on a critical mass of cars to provide enough data, and the more cars providing data, the more accurate the predictions become, BMW says. It’s currently using cars in the DriveNow car-sharing service as the main source of information for Dynamic Parking Prediction.
BMW says the i3 equipped with the prototype system successfully completed testing in Munich in March, and that the system’s “self-teaching” capabilities mean it could be easily adapted to other cities.
The i3 was chosen because it was designed specifically for city driving, the environment where predictive parking will be most needed, and because its infotainment architecture allows for greater use of connectivity features, BMW says.
Dynamic Parking Prediction also dovetails nicely with the experimental self-parking technology BMW demonstrated using an i3 at CES earlier this year.
The Remote Valet Parking Assistant is capable of driving a car in and out of a multi-story parking garage, at the whim of a smartwatch app-wielding human operator. After being dropped off, a person could remotely schedule a time for the robo-valet to pick them up.
Both technologies are still in the prototype phase, and BMW isn’t discussing plans to offer either to the public.
It’s worth noting that Dynamic Parking Prediction would be much easier to implement than anything involving autonomous driving, and could cut down on the traffic and fuel use that comes with searching for spots.
However, it also relies on a large network of vehicles providing data on available spaces, and its effectiveness will be directly tied to the size of that network.
