General micromobility giant Lime is finally rolling out some of its own Advanced Driver Assistance System (ARAS) technologies. At the Lime event in Paris, the startup shared plans to pilot its own computer vision platform that will use cameras to detect when users are driving on the sidewalk. While it will be up to the cities to warn both riders out loud about their infractions. as well as actually slow them down, both features are available.
Lime will test the technology on about 400 scooters in San Francisco and Chicago from early to mid-August. By the end of the year, Lime hopes to expand its pilot project to six cities, including Paris, where the company held a technology demo on Wednesday.
Because it is easier for cities to accuse micromobility companies and scooter drivers of riding on the sidewalk than it is to invest proper time and money in building protected bike lanes, almost every major operator has implemented some form of ARAS for scooters over the past year.
Bird, Superpedestrian as well as Neuron rely on location-based systems to determine where scooters ride and where they park. As well as voi, Rotation and Zipp have tested third-party computer vision technologies, Drover AI and Luna. Lime said it also trialed third-party computer vision systems to test the viability of the business model before investing in its own system, which will be the first camera-based sidewalk detection platform built by an operator. However, this is not the first company to integrate such a system into a scooter. Segway, which supplies many micromobility companies, recently announced its own artificial intelligence scooter.
Both camps of ARAS for scooters – localization and computer vision – have their champions. Proponents of localization say their technology is cheaper and easier to implement. today, while computer vision technology simply doesn’t exist yet to be effective. Not to mention it’s expensive and adds another piece of hardware that can break or be vandalized on the street.
Joe Kraus, president of Lime, says the company is making a long-term bet by investing in computer vision, which he believes will end up being cheaper than advanced GPS while also providing more applicable use cases.
“I’m thinking about where a lot of money is being invested to bring down the cost curve and raise the opportunity on the innovation curve,” Kraus told TechCrunch. “The amount of investment going into getting camera-based systems to do more incredible things like image classification and detection accuracy is huge. Huge investments have been made in open source software as well as chipsets to make AI chips at the edge much cheaper and more power efficient.”
In order to improve the localization-based ARAS scooter, getting ultra-accurate maps is critical to pinpoint exactly where the sidewalks actually are. Kraus does not believe that the same level of demand, investment and performance curve holds for GPS signals.
“The cost of supplementing urban GPS with more accurate timing signals is not cheap,” Kraus said.
GPS also doesn’t open up as many potential opportunities for new features, the executive claims. Lime is starting with sidewalk detection, but plans to expand into other use cases, such as parking detection. Currently, Lime and Bird are testing the Google ARCore Geospatial API. which uses the driver’s smartphone camera to accurately geolocate parked scooters and e-bikes to determine if they are properly parked.
Lime may also decide to use its computer vision technology for “broader localization efforts” or “non-repudiation in the event of accidents,” Kraus said. Lime Vision, as the feature is called, can also help Lime win brownie points with cities by sharing data about things like pavement driving hotspots to help know where to actually build bike lanes, or the number of potholes on certain streets.
The first version of the Lime Vision to be tested with an initial pilot project is an upgradeable waterproof device that attaches to the scooter’s neck below the handlebars and contains a camera, an AI chip, and a CPU or processor. This is where the Lime computer vision model performs its calculations in real time, Kraus says, allowing objects to be detected in less than a second. The device is wired to the scooter’s brain so it can send commands based on what it analyzes.
According to Kraus, Lime is already working on a second version of its computer vision system that will be fully integrated into the body of the car, rather than being a stand-alone attachment.
In addition to the computer vision launch, Lime said on Wednesday it plans to pilot a reaction test to prevent drunk driving in 30 cities around the world. The test, which will appear in the Lime app when a driver tries to book a ride after a designated time of night, looks like a game of sorts: the user drives a scooter as he drives down the street and is asked to press a button. stop sign when it appears. Riders who do not click on this stop sign within a certain period of time will not be able to start the ride.
Lime previously tried out a similar feature that activated after 10 p.m. in most markets. It will ask drivers to enter “YES” in response to the question: “Do you confirm that you are not drunk and cannot drive?” Last year, Bird launched Safe Startan in-app checkpoint that asked drivers to enter a keyword into the app that they hoped would keep drunk people from driving.
Credit: techcrunch.com /