EV charging sucks — can smart cameras make it better?

Public EV charging often involves juggling a variety of apps, each with its own bespoke sign-up process, and attempting (and often failing) to tap or swipe a credit card to pay for the charge. Occasionally, there’s a QR code involved. Overall, it’s a pain in the ass — and an obstacle to widespread EV adoption.

Now, one EV charging operator is betting that computer vision can help take the stress out of it. Revel, the shared moped-turned-ridehail service, is partnering with a startup called Juice to install smart cameras at its EV charging sites in New York City that are mostly utilized by rideshare drivers. By using computer vision, Juice can identify the model and license plate of the vehicle in need of charging and, then after an initial sign-up, will initiate the charge without any additional work by the driver. No apps to download, no juggling various credit cards. Just flowing electrons. 

“Our goal is not to create a long experience where we’re forced to come in and download an app and use that app every single time,” said Paul Suhey, cofounder and chief operating officer at Revel. “We’re just trying to make the experience as easy as possible.”

In some ways, Juice is trying to replicate the experience that many Tesla owners have enjoyed for years. One of the biggest selling points for Tesla is its extensive and reliable Supercharger network. Tesla owners input their payment information in the Tesla app once, and then, after pulling up to a Supercharger and plugging in, the company takes care of the rest. 

Unfortunately, the rest of the public charging network is not so seamless. Major operators like Electrify America, Charge Point, and EVgo, route much of the functionality through their respective apps. Often, it works fine, even if there are a few extra steps. But when the communication between the hardware and software is less than reliable, the experience can be incredibly frustrating. EV owners frequently cite glitchy software as one of the main culprits in subpar charging.

Revel is dealing with more than just Tesla drivers at its EV chargers, even though its rideshare fleet is made up of mostly Model Ys, as well as some electric Kias. Its EV charging stations are open to all models, offering plugs for both Teslas (SAE J3400, or NACS) and non-Tesla vehicles (CCS1 and CHAdeMO). As such, Revel needs to level the playing field for all the Ford, Hyundai, and Kia EVs as well. And that’s why it turned to Juice to help smooth out the bumps in the road. 

James Murfin, cofounder and CEO of Juice, told me he got the idea after staking out several public EV chargers in the US and Europe with a stopwatch and clocking how long it took for people to initiate a charge. 

“We just got so frustrated with all the user experiences,” he said, “different apps, different QR codes, different connectors, different charges, some working, some payment terminals broken.”

He wanted to create something that was analogous to Apple’s Face ID, something that just works reliably after an initial signup process. Like your face, your vehicle has a lot of unique identifiers, like its paint color, model type, and license plate number. 

To sign up, all you need to do is scan a QR code — yes, I know — which opens up a webpage, no app download necessary. From within the webpage, Juice asks for permission to access your phone’s camera to record your vehicle from multiple angles, including your license plate. The only other step is to choose your payment option. Most people will likely choose whichever credit card they already have stored in their phone, either through Apple or Google Pay. 

“And then essentially you’re done,” Murfin said. “You never have to get your phone out again. You drive into that Revel station, plug in, juice up, and you go.”

In terms of the hardware, Juice has two methods: either they can use existing security cameras and “block” each stall using machine learning so they can identify charging vehicles, or they can install pinhole cameras at each charger. Once a car pulls into a charging space, Juice’s machine learning is already at work identifying that particular vehicle, and initiating a charging session — even before the driver gets out of the vehicle. 

On the backend, Juice is collecting data on each charger session to send back to Revel, so it can see how EV owners use its chargers. Most of Revel’s customers are rideshare drivers who are just looking to get in and get out, as quickly as possible, Suhey said. And they charge a lot more frequently, often two to three times a week, than normal EV owners.

“So as a component of the experience, they were definitely interested in how you can pull up and just make it as frictionless and as seamless as possible,” Suhey added. 

And Revel is already swimming in data from its customers. Since integrating Juice technology at its three charging sites last month, the company says it has already logged more than 6,000 rideshare driver sessions, with over 1,100 drivers active on the platform. Dozens of drivers have done more than 20 individual sessions in a month, which Revel argues is evidence of Taxi and Limousine Commision-licensed drivers being a “super user” group for charging sites.

Moving forward, Revel has plans to install around 200 additional EV chargers in New York City, including 60 in Queens, 48 near LaGuardia Airport, and more in the South Bronx. The city’s TLC recently released 10,000 new rideshare licenses just for EVs, bringing the total number up to 12,000. Revel operates its own fleet of 550 Tesla Model Ys; the rest are independent Uber and Lyft drivers or affiliated with black car companies. Last year, Revel’s chargers were mostly being used by its own rideshare drivers; now, it’s about 50-50 between its drivers and independent app drivers. 

All of which is to say, Revel and Juice have a lot of work ahead of them. 

“One component of how we can use this computer vision technology is really just to understand our customers more,” Suhey said, “to inform our decisions on where we’re acquiring sites, where we’re developing, [and] how we’re designing these sites.”