Electric scooter operator Superpedestrian has acquired Navmatic, a startup that helps micromobility operators locate vehicles and correct their movements in real time.
The companies did not reveal the details of the buy, which was finalized last month. Those close to the deal say the number is not far off from Navmatic's last funding round, which was $4 million raised in seed funding last June.
The Navmatic purchase means Superpedestrian can apply the startup's Super Fusion technology to enhance its vehicle safety systems. Pedestrian Defense, as the new system is called, can detect unsafe riding behaviors -- like riding the wrong way down a one-way street, aggressively swerving, sidewalk riding or repeated hard braking -- and either notify the rider or correct the rider’s behavior in real time by slowing or stopping the scooter. Riders receive a safety rating at the end of the ride that is used to deliver customized safety training, to incentivize good behavior via discounts or to blacklist chronically unsafe riders.
As more stories surface about scooter-related accidents, picking on the shiny, new tech amid the commonplace majority of car-related road collisions, Spin and Voi have also tacked on additional tech meant to keep riders in line and protect pedestrians. Rather than focus on positioning software like Navmatic’s, they’ve turned to computer vision startups Drover AI and Luna. Given an unobstructed camera, Spin and Voi vehicles can, unlike Superpedestrian’s, actually detect pedestrians with some degree of accuracy. The difference that Superpedestrian is offering is the ability to pick up on precise rider behavior via data on the scooter's micro-movements.
“The current challenge is how to protect the more vulnerable people around a scooter, meaning pedestrians, people with disabilities, a man or a woman with a stroller,” Assaf Biderman, CEO of Superpedestrian, told TechCrunch. “For that you need a very precise location, you need characterization of the rider's behaviour and you need context awareness, like is the rider blocking the right of way? You don’t need a camera if you train your data right.”
Biderman says Superpedestrian runs a microprocessor with Navmatic’s software on its LINK scooters' operating systems, which is where all of Superpedestrian’s maps live. That software is trained on a variety of sensors, including vision for ground truth and high definition maps. During a ride, real-time computations analyzing data picked up from the scooter are done on the edge of the microprocessor, combining raw GPS data, multiple dimensions of inertial sensing and vehicle dynamics to calculate very precise location and movements of the vehicle.
“With Navmatic’s software, location detection significantly improved, and all of a sudden you can really do this kind of analysis of the minute motions of someone on a scooter or small vehicle,” said Biderman. “Our response time is now about 0.7 seconds.”
According to Paul White, director of development and public affairs at Superpedestrian, the vehicle’s location accuracy improvement ranges from 70% to 90% depending on local conditions.
Both companies say such precise location data, coupled with the ability to control the vehicle’s movements, is a superior solution to vision, one that is potentially cheaper but definitely more scalable.
"You're not gonna put a $1,000 or $2,000 lidar on a scooter, right?" said Biderman. "What you're able to do through sensor fusion helps you to overcome the limitations of vision alone, which is bad at night, can get dirty and can fail a lot, or GPS alone, which can suffer from reflection and shadows. When you combine as many sensors as you can, you benefit from the best of all worlds, and it just keeps learning and improving."
For reference, tech from Drover AI and Luna also allow the scooter operator to control the rider's movements based on data it picks up, but those capabilities aren't yet live in every city in which Spin and Voi are using computer vision. While Navmatic's chip shares an operating system with the scooter, Drover AI's for example lives on a separate onboard IoT unit that communicates directly with the scooter's OS. This capability is being piloted by Spin now in Santa Monica. So far Voi has just piloted Luna's technology in Cambridge with audible alerts to riders on the sidewalk, but is exploring the option of slowing the scooter down.
Superpedestrian has boasted its ability to better control its vehicles by owning the full stack, from operating system to hardware, compared to other operators that buy off-the-shelf operating systems.
"We had a lot of challenges trying to integrate our tech with other companies because every time they needed to change a line of code, they had to contact the manufacturer, which could take a week to make any changes," Boaz Mamo, CEO and founder of Navmatic, told TechCrunch.
Superpedestrian's software-first approach to this acquisition was also attractive to one of the company's current investors, Dan Herscovici, partner at Edison Partners.
"Most other acquisitions by scooter companies are usually a play for market share," he said. "It's not as typical to see micromobility companies really leaning on IP and tech to really enhance the vehicle."
Herscovici said Superpedestrian looked at many solutions in the marketplace to solve for the problem of pedestrian safety and city compliance, and even considered developing its own tech. Balancing time-to-market and the need to move quickly, the ever-present threat of losing city permits looming over their head, acquiring Navmatic seemed like the right decision.
"The way I generally like to think about the micromobility space is that there are three major constituencies," said Herscovici. "The first is the rider, and that's what most micromobility companies think about. How do you attract them and keep them riding? The next is the city or municipality, and the next and most forgotten about is the non-rider, the ones sharing the road. I think the industry has been searching for a way to modify rider behavior based on safety regulations and this acquisition unlocks that capability."
Mamo also pointed out that Pedestrian Defense unlocks the capability for cities to get some insights into how riders are traveling and how they might be able to make better infrastructure decisions.
Biderman said Superpedestrian is building around 50,000 scooters this year, with every new vehicle the company builds containing this new technology starting in December. Next year, the company will begin deploying the upgraded vehicles in the new cities it's moving to, as well as switching out older models in cities where Superpedestrian has an existing presence.
This article has been updated to reflect correct information about the capabilities behind Drover AI and Luna technology.