Design Engineering

Self-driving cars could make urban parking easier


General Automotive U of T

A UofT Engineering study shows that autonomous vehicles can be packed into tight parking lots and rearranged when signalled.

U of T parking autonomous vehicles

Left to right: Sina Bahrami, Mehdi Nourinejad and Professor Matthew Roorda designed an algorithm to optimize the design of parking lots for autonomous vehicles, increasing their capacity by an average of 62 per cent. Photo courtesy of Robert Baker.

As more and more automakers are jumping on the autonomous vehicle bandwagon, researchers are still trying to determine the implications of this technology. A new research study a U of T Engineering believes that adopting self-driving tech could reduce urban space needed for parking.

Self-driving cars can be packed into a parking lot tightly in a way that regular human-driven vehicles cannot.

“In a parking lot full of AVs, you don’t need to open the doors, so they can park with very little space in between,” says Professor Matthew Roorda, senior author of a new study in Transportation Research Part B. Roora adds that you don’t even need to leave space for vehicles leaving first, as the surrounding vehicles can move out of the way when signalled.

The goal of the study was to determine the optimal size of the grid to maximize storage while minimizing the number of moves required to extract any given car.


The team determined that an AV parking lot could resemble a solid grid structure rather than islands of cars.

Mehdi Nourinejad, a recent PhD graduate from the Department of Civil Engineering and the study’s lead author, explains that there are some tradeoffs with this configuration. A very large grid could lead to a lot of relocations, which would make getting vehicles out much longer. On the other hand, a lot of space is wasted with a number of smaller grids.

Nourinejad, Roorda and their co-author Sina Bahrami created a computer model to simulate the effects of various layouts for AV parking lots. They then used an algorithm to optimize the design for various factors, including minimizing the number of relocations and maximizing the proportion of the lot that was used for parking versus lanes for relocation, entering or exiting.

A well-designed AV parking lot could accommodate 62 per cent more cars than a conventional one, making better use of limited space. However, this only works if the lot is designated solely for AVs.

Another advantage to this type of lot is that it is not fixed and can be modified on-demand to accommodate changes. Bahrami explains that you don’t need to paint new parking spaces, but rather signal vehicles to move about, fitting more cars in.

This new paradigm could also introduce negative consequences, such as a potential increase in traffic congestion.

Roorda and his team also can’t predict when the number of AVs on the road will reach the critical mass required to make use of their designs.

“We’re talking about large numbers of vehicles that can fully drive themselves, with no requirement for a driver to take over if something goes wrong,” says Roorda. “There’s a lot that has to happen before we get to that stage.”


Stories continue below

Print this page

Related Stories