TIER Brings Edge Computing to Its Micromobility Fleet on AWS IoT

Since 2018, there has been an explosion of on-demand electric scooters in cities around the world, representing a significant, ongoing transformation in personal mobility. However, as long as steep maintenance and support costs keep micromobility companies from pushing innovation further, that transformation will remain incomplete. This is why Berlin-based micromobility company TIER Mobility (TIER)—whose mission is to “change mobility for good”—has broken out of the standard maintenance model, having discovered a way to drastically reduce operational costs while making personal mobility more accessible and affordable.

At the center of TIER’s strategy to reduce costs by improving and streamlining battery swapping and recharging for its fleet of scooters is Internet of Things (IoT) technology powered by Amazon Web Services (AWS), specifically AWS IoT Core and AWS IoT Greengrass. These services have the potential to reduce TIER’s battery-related operating costs by a factor of 10, creating savings the company can in turn pass along to customers. Beyond that, the AWS-powered IoT environment presents TIER with further opportunities to quickly develop viable micromobility solutions involving weather detection, speed restrictions, traffic information, and more—all while maintaining control over its data and brand experience.

Since launching in Vienna 2 years ago, TIER has expanded to around 80 cities in Europe and the Middle East. It has honored its mission to change mobility for good by becoming the first climate-neutral company in its sector and by offering helmets that fold into its scooters and equipping the steering bar and back fender with indicator lights for rider safety. But TIER needed to solve another challenge. Put simply, the traditional model of operating and maintaining scooters costs too much. Scooters are intended to be scattered across a city, making centralized service depots expensive. What’s more, scooters require constant attention. “They don’t last long if you don’t treat them as well as we do,” says Amir Hadi, TIER’s director of technology. When in high demand, scooters also may need to be recharged every day. The recharging solution that other ride-sharing companies employ—gig workers pick up and recharge scooters low on battery—is itself expensive and precarious, and frequent loading and unloading can cause undue damage.

To solve the charging issue, TIER did what it does best—imagined a better complete customer experience and economic model, which in this case involved reinventing scooter monitoring and charging. The company turned to its usual method of problem-solving: “Our strategy is to always use off-the-shelf solutions where good solutions exist and identify the areas that we want to take ownership over,” explains Hadi. TIER found its off-the-shelf solution on AWS IoT Core, the managed cloud service that lets connected devices easily and securely interact with cloud applications and other devices, and AWS IoT Greengrass, which extends the capabilities of the cloud into edge devices. For TIER, AWS IoT Greengrass could enable each scooter to act locally on the data it generates while still using the cloud for management, analytics, and durable storage.

TIER had previous experience with off-the-shelf AWS services—namely, Amazon Elastic Kubernetes Service (Amazon EKS), the fully managed Kubernetes service AWS offers. “All our backend services run on Amazon EKS,” says Hadi. “Our developers enjoy it because they can develop the code locally and run their container locally. They just push it to production, and it’s basically the same execution environment as on their computers.” Just as Amazon EKS offered an off-the-shelf solution for backend development, so could AWS IoT Core and AWS IoT Greengrass facilitate innovative solutions for TIER’s maintenance challenges.

When TIER acquired Pushme Bikes in early 2020, the company acquired a prototype charging station that can recharge multiple batteries simultaneously. “It was a really cool solution,” says Hadi, “but we needed to enable backend developers to run code on the edge. So we basically switched the IoT part of the solution and deployed AWS IoT.” To avoid latency challenges that often come with more centralized IoT infrastructure, TIER opted to include AWS IoT Greengrass in its solution. “Normally you need quite a lot of knowledge about embedded computing, but on AWS IoT Greengrass, we can take it to a higher level and help our developers contribute logic for the edge. We also don’t want to calculate too much on the cloud,” says Hadi. “It’s far away from our customers. We want to calculate and compute all things as close as possible to our customers—that’s why we chose AWS IoT Greengrass.”

Edge computing would enable the devices themselves to monitor data and make decisions locally—like unlocking to begin a ride—without being overly reliant on the cloud. “Edge computing puts customers at the center of our operations, which would enable them to always use our vehicles even if connectivity is down,” says Hadi. With the aim of having this technology ready by the beginning of 2021, TIER will use what Hadi calls “after-the-fact architecture” so that customers don’t miss a ride just because Wi-Fi has dropped. Instead of waiting for the command to come from the cloud, the e-scooter would unlock, let the customer ride, and bill the customer once the connection is restored—sometimes even after the ride is completed.

AWS IoT Greengrass and edge computing were behind TIER’s first solution to reduce operational costs—a battery-swapping system that combines two kinds of hyperlocal intelligence. One is a sensor on the scooter, which determines when batteries should be exchanged. “The device itself constantly monitors all kinds of metrics, like voltages and temperatures, and then tells our backend when a battery can be swapped,” says Hadi. The other is TIER’s community of riders, who receive information both before and during the ride if a scooter is eligible for an incentivized swap. Riders can then reduce their costs by volunteering to swap out batteries at a charging station.

Importantly, the device is not constantly sending all battery data to TIER but sending a “smart update interval.” In other words, a small change in battery temperature will be monitored locally but not communicated to the main server. “A larger temperature rise, say 5 degrees, would prompt the edge-computing capabilities to say, ‘OK, this is worth an update to the cloud,’” says Hadi. This feature of edge computing reduces network traffic by letting the device determine what is important enough to share. The smart charging system is further improved by access to AWS IoT Analytics, which quickly alerts TIER’s on-call team to any serious issues. “If we see a drop in one of the key metrics, we get an alert,” says Hadi. “The on-call team sees what the problem is and fixes it very quickly.”

Empowering riders to help with battery charging through an innovative battery technology that was introduced with the latest generation of TIER scooters has the potential to save the company a considerable amount on operational costs by avoiding the old method of picking up scooters to recharge their batteries. TIER chief technology officer and cofounder Matthias Laug expects the battery-swapping program to reduce battery-related operational costs by a factor of 10. “This results in all kinds of side benefits for the customer, such as better prices, more scooters, and higher availability,” says Laug.

TIER’s AWS-powered IoT solution has a number of implications in terms of crash recovery and the development and deployment of new features. “If we have a major outage, usually it takes us a couple of minutes to understand and fix the problem on the backend,” says Hadi. “So our backend goal—and it’s a very conservative one—is to recover fully within 2 hours.” Hadi says the team is shooting for 50 software deployments per week per team once the functionality is up and running across the TIER fleet.

With help from AWS, TIER has found a way to reduce operational costs while delivering a better experience for riders. With new IoT infrastructure in place, TIER is poised to deliver a wide range of customer-centric new features. For example, in the AWS IoT Greengrass environment, TIER could add functionality based on where scooters are operating, including through the detection of weather, speed restrictions, traffic, or other events that would change trip time. And the near-zero latency means customers could eject batteries instantly or slow down just when they enter restricted speed zones. “The dream here is to have customers use our vehicles without taking their phone out at all,” says Laug. “That’s the vision. You jump onto the vehicle, and the vehicle’s on. This is what we’re aiming for.”

To learn more, visit aws.amazon.com/iot/.