Toronto Toll Road Takes the Fast Lane to Personalized Service
The 407 ETR, Toronto’s all-electronic, barrier-free toll highway, is just an example of the capabilities in a connected world, according to Ravi Chandler, group architect for the project.
Speaking at ProgressNEXT 2019, the Progress Software global user conference, he outlined three scenarios in which users could seamlessly use its mobile app:
- You’re headed to an interview for your dream job.
- Your wife goes into labor, but isn’t going to make it to the hospital to deliver.
- A friend from another country comes to visit you.
In the first, if you’re running late, and you find the fastest route is to take the 108 km (67 miles) toll road. You can just take it without signing up or looking for correct change.
If you have to pull over on the toll road with the baby coming, with one click, you can summon help with emergency vehicles that patrol 24/7. The app will pinpoint your exact location, notify you who is coming and how long it will take for help to arrive. If it’s not an emergency, say you just run out of gas, the emergency crew will bring you enough gas to get to the next gas station at no charge.
Through the app, your foreign friend also can authorize you to pay her toll, so she doesn’t have to worry about getting a device for her car or exchange rates. When she exits the tollway, the app will notify you that she took the correct exit.
In developing the app, the group set out to create something completely different, Chandler said.
“We don’t see other toll highways as our competitors. Our competitors are any product that delivers a remarkable customer experience,” he said. Like Uber or Amazon Prime, he said, the goal is exceptional ease of use.
The privatized ETR (Express Toll Route) has electronic sensors on overhead towers that log each driver’s entry and exit point that can identify the vehicle from a transponder signal and/or take images of a license plate even at vehicle speeds greater than 200 km/h (124 mph).
A lot of data has to come together to make this happen from disparate backend systems and heterogeneous technology stacks to ensure it is processed in real time. It had to use modern application development processes and it had to scale.
Doing this themselves would have required a lot of overhead and a ton of code, Chandler said. Instead, it needed a platform that could combine device data and backend data.
“[Progress] gave us a very clean way of focusing on our value proposition without having to worry about all these solutions. We found it very opinionated in certain ways, but we found we agreed with those opinions more often than not, and when we wanted to tweak it, we could,” he said.
The ETR already was using Progress tools for managing its website and business rules — for instance, it sets rules for when and how often to contact drivers about delinquent payment before beginning a complicated process that the province enacted to deny license renewal to vehicles with outstanding bills. That’s a last resort, Chandler said.
The two biggest challenges in creating the mobile app, he said, have been accuracy and latency.
“If we don’t want people to have to have transponders, we’re relying on GPS technology,” he said. “GPS technology can be incredibly accurate, but it’s not 100 percent accurate. How to we achieve that last bit of accuracy to say you really got on at this exit and you didn’t swerve away at the last minute? [We use] smoothing out algorithms and pattern recognition to say they really did get on there and got off there.”
The toll system uses transponders and Optical Character Recognition (OCR) technology today, which comes with a margin of error. Adding new technology adds another margin of error. So how do you know you’re not compounding errors? What you’re actually doing is complementing the data itself, he said.
“What we wouldn’t want to do it say this is only the mobile device trip or only the transponder trip or the completely device-free trip. We’d want to amalgamate all these vectors and use them as input,” he said. “Take the best of each feature and apply that to accuracy. Accuracy is the biggest technological challenge, and you mitigate that by using the most inputs as you possibly can.
Latency poses another problem.
“[Everything] has to go through several integration layers. We have to get your location data, apply algorithms on it, apply a pricing model on it — we have to a bunch of different things through different integration layers both in the cloud and on-prem before it ever gets to you. Right now, we can get it down to around two seconds, which is an enormous technological achievement, but it means performance optimizing every single layer,” he said.
Security and authentication were basic requirements that had to be baked in from the start. The service is mindful of sensitive information — and the potential creepiness of monitoring your movements, he said, and uses the least information needed to complete each function.
Much has been made at the Progress conference about the need to create personalized experiences. But how do you personalize a toll road?
“We can figure out that Monday through Friday, nine to five, you go to work. On Saturday, you take your kid to soccer practice. On Sunday, you go to church. We can notify you of an accident and suggest an alternate, but not be heavy-handed about it, not always be promoting the 407, but offering you the one that’s faster,” Chandler said.
“It’s that app saying, ‘You’ve used the toll road four out of five days this week, go have a free coffee on us.’ If we have a partnership with that [coffee shop] we use their APIs, pay for their coffee.”
“We want people to say, ‘I want to use this service. It’s awesome.’ That’s what we’re competing against.”
Photo via 407 ETR