How AI Will Shorten Your Commute Through the City

And track your every move.

|
Nov 14 2017, 8:47pm

Image: Guy Torsher

Forecast is a series exploring the future of AI and automation in a variety of different sectors—from the arts to city building to finance—to find out what the latest developments might mean for humanity's road ahead. We'll hear from Nikolas Badminton, David Usher, Jennifer Keesmaat, Heather Knight, Madeline Ashby and Director X, among others. Created by Motherboard in partnership with Audi.

On November 7, 2017, Toronto City Council voted 36-6 to make 2.4 kilometres (1.5 miles) of temporary bike lanes along the major commuter route of Bloor Street permanent.

The decision came on the heels of a year-long pilot project and decades of advocacy by the city’s cycling community and was, as the CBC described it, “Toronto's most-intensely studied transportation project.”

At the heart of that decision was data collected by Kitchener-based Miovision Technologies, which specializes in traffic data collection and advanced signal light operations. And increasingly, artificial intelligence is a core component of that business.

Former Toronto chief city planner Jennifer Keesmaat is an advocate for smarter decisions based around better data, she has some concerns about how that data is collected and used. She admits with all the AI-enhanced technology tracking us, some of it without our knowledge or consent, it's reminiscent of George Orwell’s dystopian classic 1984.

Miovision used what they call the Scout, a portable and weatherproof camera that uses AI and deep learning to analyze traffic in real time and feed the results to traffic planners. It’s been used by more than 13,000 cities, counties, and transportation agencies around the world, according to the company.

The city’s Bloor Street bike lane evaluation report found cycling volumes increased by 49 percent after installing the temporary bike lanes, and the average weekday volume of 5,200 cyclists ranked second in the entire city, and there was a 44 percent decrease in all conflicts along the street between motorized vehicles, pedestrians, and cyclists.

Commute times did increase for drivers, by as much as four minutes, and that drew the ire of some councillors who voted against it, but ultimately the positives outweighed the negatives in the eyes of the city.

Miovision CEO and co-founder Kurtis McBride said the use of this kind of data is indicative of the smart city movement—using new technology, including AI, to help change the ways we move around our cities.

“Every city has a different philosophy around driving, walking, transit or whatever is best. [Miovision’s role] is to not take a stance on one or the other. We’re providing technology to help cities measure the right things to drive the best outcomes for citizens,” McBride told Motherboard.

Consider the example of a double-parked vehicles in the roadway. Miovision has a contract with a “major North American city” to monitor streets using AI algorithms to look for these types of lane obstructions and help staff respond to the obstruction.

“One of the biggest causes of congestion in downtown cores is when a delivery truck pulls in for three minutes and blocks a lane, then all of a sudden it blows up the whole network,” McBride said. “Long after that truck has left, the network is still healing itself.”

And when a major highway is closed due to a crash, their predictive AI can assess how a traffic network will be impacted when cars spill into city streets to find a detour.

Keesmaat told Motherboard that data must be king when it comes to making decisions around how to plan a city and improve its transportation infrastructure.

“Once we get the data, it can transform the decisions that we make because we understand in a better way what is happening in the urban environment,” she said. Think of the classic example of adding more lanes to a road to try and ease congestion. It seems like a good idea, but statistics don’t back it up, Keesmaat said, thanks to a phenomenon known as induced demand.

“Which means in a city or region where there is more demand than traffic capacity, any time you add traffic capacity it is quickly absorbed,” she said. It’s the classic example of buying a bigger belt to deal with your weight gain, instead of going to the gym.

“This is why data is really critical, and collecting that data in a robust manner is important for making good decisions that are in the long-term interest of city planning,” Keesmaat said.

Miovision also deploys a permanent traffic monitoring system called Traffic Link. It uses cameras and Internet of Things connectivity to scan intersections and provide real-time data about how traffic is moving. Deep learning AI helps determine how vehicles move through the system in real time, such as how long certain vehicles wait to make a left-hand turn at an intersection, and traffic engineers can adjust the timing as needed.

Keesmaat said there is a place for government regulation in AI, much like there is in other areas of city planning.

“What’s clear is there is so much data being collected on us all the time, in some ways it’s very creepy, the question is how much it begins to inform our political process and our policy making,” Keesmaat said. That includes increased government regulation over the development of AI and its safe deployment in a society where very few of us understand just how much data is being collected about us every minute of every day.

“We don’t expect people who buy a condo to know what makes a safe building, we expect the government to do that through the building code,” she said.

McBride said as a vendor hired by a city or transportation service to provide data, all that data belongs to the client and is not sold or redistributed. It’s stored in the cloud on the Amazon Web Service Cloud and has some of the best security in the world, according to McBride.