Crowdsourcing Road-Quality Info With a GPS Phone App

November 15, 2019
It was just past 2 a.m. on the leafy streets of Lexington, Massachusetts, when Jacob Roxon found himself abruptly blinded by a flashlight blaring through his driver's side window. He'd been sitting in his car on the side of the road, staring at three phones on the passenger seat, when an officer cruising by, pulled up, and asked why he was hanging around the darkened homes of this suburb outside Boston.
"I told them I was taking road-quality measurements," Roxon, a PhD student at the MIT's Concrete Sustainability Hub, later told Cheddar. "They said, 'Huh, that's interesting, considering that that road had recently been resurfaced.' They found it hard to believe."
Such are the hazards when test-driving a new app – Roxon had been cruising the empty streets at night for months, sometimes driving hundreds of miles a night in rented cars and SUVs to put the latest versions of the app, Carbin. He and a team of some half-dozen researchers even created miniature traffic jams on MIT's campus to simulate how a short-lived backup can create waves of highway congestion for hours.
Their goal: To develop an app that tracks the road's quality while you drive – and, one day, offer GPS navigation that points the way to not simply fastest route, but also options for the smoothest, cheapest and, most of all, greenest way to get there. The name "Carbin" is a mashup: Car for, well, "car," and "bin," for "binning carbon" – as in putting carbon in the trash bin.
"Most of the time people think that fuel consumption comes from how big the vehicle is, the type of engine that you have, and the way you drive. And it's true, the majority comes from that. But a lot of it also comes from road quality, from something that drivers don't even think about: Once you're on the road, you just take the road that's fastest," said Roxon, who hails from the United Kingdom and is overseeing the project. "If we are able to identify roads where you have lower fuel consumption, we're going to be able to reduce carbon emissions."
Bad roads in a city can drive up fuel consumption – and increase greenhouse gas emissions – by as much as 15 percent, according to data that Roxon and his team have collected. On highways, where average fuel economy is higher, rough roads can sap 5 percent from fuel economy. But even that figure can vary lane-to-lane, the Carbin team found.
A weathered right-hand lane worn down by heavy-duty trucks and buses can generate 2 to 4 percent more heat-trapping emissions than just the next lane over. Combined with the wear and tear on the car, the cost-per-mile can jump as much as 10 percent from one lane to the next – perhaps a small consideration for the average driver, but big bucks for long-haul truckers, delivery companies, cab hacks, and others who spend their workdays on the road.
"If you start to think about it on the aggregate scale, the city, state, or national scale, then 10 percent savings translates to all the sudden a lot of money for people, and it translates to significant CO2. That's where we see the future," Roxon said.
The climate implications are also considerable: "If you have a lot of heavy-duty vehicles driving in that particular lane, the impact is going to be more severe than if you had the same number of passenger vehicles. Diesel emissions are far worse than gasoline emissions – and all the heavy-duty trucks use diesel engines."
Carbin's main interface is GPS navigation, built on software from a third-party maps and GPS provider, Mapbox. But with one button-press from a user, Carbin will start tracking and recording road roughness by using the accelerometer built into nearly every smartphone. The more the phone gets shaken about, the worse the road.
After each trip, the app beams the results to Carbin's servers. After getting data from the same road – or even the same lane – from three to five users, they'll analyze the findings and upload them to, a color-coded map of road quality around the world.
The typical department of transportation can take as long as two years to measure a single length of road – a process that requires a $250,000 laser-measuring device, driving the assigned stretch at a constant speed, only when the road is dry, and repeated trips if there are multiple lanes.
By contrast, Carbin and its road-tracking website, FixMyRoad, is crowdsourced in near-real time. From the red and orange veins tangled at right angles across central Boston and D.C. and Los Angeles, to the greener stretches across the suburbs, the results for the aggrieved city driver are at once deeply validating – public, empirical proof that the roads are, in fact, as bad as they seem – and promise no shortage of fodder for city planning meetings.
"What we're interested in doing is finding, understanding, how the performance of the road changes over time. This is something that the government doesn't do well," Roxon said. "You need to predict where and when a given road segment is going to need replacement, so you can plan efficiently where to make repairs."
Roxon and his team hope that by scoring road segments on a color-coded spectrum from green to red, they'll help cities, counties, and states better decide where to invest in road improvements. But the predictions aren't limited to streets and highways – they think they'll be able to apply it to drivers' own vehicles, too, essentially by measuring its heartbeat.
"Based on the vibrations, we're going to be able to figure out if it's a four-cylinder engine, if it's an electric engine, if it's a V8 engine, because each one has a different characteristic signal of vibrations," Roxon said. The same goes for worn tires, axles, shocks, and more. "With the app we're going to be able to provide the user about the quality of their vehicle: When is the right time to make repairs?"
The Carbin app officially went live in February, the product of students and faculty at MIT, the University of Massachusetts Dartmouth, the American University of Beirut, and Birzeit University in the West Bank. With data from the run-up to the launch and the months since, it's racked up more than 160,000 miles in road trips around the world, many of them around Boston. It boasts about 400-500 active users, who have submitted 3,000 recorded trips.
The road from novelty to full-on success, though, still stretches a ways further: Carbin's developers estimate that they'll need to log one million miles and recruit roughly 5 percent of a region's drivers to consistently use the app for Carbin to start monitoring – and begin making predictions – about a road network, its traffic density, and even each user's own vehicle. In Boston, that represents about 75,000 of the roughly 1.5 million motorists who commute throughout the metropolitan area.
The team hopes to use a points system – like in Waze or even with airline miles – to incentivize users. It's seeking about $500,000 to $1 million in funding for the next stage of development.
"Most of the drivers in most of the countries around the world already have smartphone devices where they're already using Google Maps or Waze," Roxon said. "They're already using their phones while driving, so why not use the phones to do something extra that would be helpful?"
We use cookies and similar technologies on this site to collect identifiers, such as IP address, and cookie and device IDs as described in our Privacy Policy.