Our mission at Spin is to create a world full of 15-minute cities—cities where everything people need is easily accessible within a 15-minute scooter ride, bike ride, transit ride, and/or walk.
Sidewalk riding and parking issues are two of the top complaints we hear from city leaders and members of local communities. It’s the driving force behind why we partnered with Drover AI to launch Spin Insight Level 2.
Technology is not the only way to resolve the problem; we need infrastructure such as bike racks, scooter corrals, and bike lanes to make space for not just e-scooters, but other two-wheeled and three-wheeled devices on our streets.
However, if we want micromobility to continue flourishing in more cities around the world, we have to take an honest look at the issue to 1.) help cities make informed decisions about where they should invest in infrastructure and 2.) develop technologies to ultimately change the behavior of our riders. Right now, we are changing behavior using our Spin Insight Level 2 technology powered by Drover AI in five cities: Santa Monica, California; Atlanta, Georgia; Miami, Florida; Seattle, Washington; Milwaukee, Wisconsin.
Just like the hype and swirling misinformation surrounding autonomous vehicle technology capabilities, each company in our industry is also using marketing and press to ensure regulators view their solutions as the best (even if that technology is proven to not work). That’s why we are taking the time to lay out the facts behind all of the solutions currently on the market right now.
Spin Insight Level 2 powered by Drover AI uses a camera and machine learning artificial intelligence to “see” a rider’s surroundings and make decisions in real-time. Camera-based solutions use forward (and sometimes backward) facing cameras, along with various detection algorithms, to determine what is in front of or around the vehicle. This information can be used to identify city infrastructure such as sidewalks, bike lanes, parking corrals, and curbs and in the future can also be used to identify pedestrians or obstacles in a rider’s path.
Just like your brain needs your eyes to see what’s around you and react to your changing environment, camera-based technology is necessary to quickly and accurately understand what is around a vehicle and react in real time. Without your eyes, you may, with time, become comfortable in your own home, learning and memorizing where the furniture is by touch. However, as soon as you leave your home and enter more complex and dynamic environments with other people and vehicles, your understanding of your surroundings becomes much more unclear, and contextual awareness is critical. That’s why a camera is such an important part of the equation when developing technology that effectively combats sidewalk-riding, bad parking, and right-of-way violations while accounting for advancements in technology that will protect riders in the future.
Possibly the most important thing to know about camera-based technology is that it has the capability to improve over time and stay up to date with advancements in artificial intelligence. AI and machine learning technology are two of the fastest moving sectors in all of tech. Advancements are being made so rapidly that the capabilities of AI and machine learning in a few short years will look much different than they do today. This should be considered when making decisions about what technology will be the most effective at combating common micromobility complaints. For example, future on-board camera technology could very well be able to prevent collisions before they happen, similar to advanced driver-assistance systems in personal cars. The method cities choose should solve the problems of today and be capable of taking advantage of new and improved technology over time.
In addition to AI advancements, there will also be new ways to use the images analyzed by the camera. Camera technology will have a major impact on rider safety even outside of sidewalk riding detection. With the amount of data gathered by on-board cameras, it will be possible to use it to better understand safety incidents and why they happen, instead of just where they happen, which is all GPS technology can tell us. On-board camera tech is used similarly in this study conducted by Virginia Tech researchers. While this technology isn’t being implemented at scale today, it will require a camera when it comes to fruition in the coming years.
Another tool used by micromobility vendors to combat sidewalk-riding is location-based technology. High Precision Location (HPL) technology can improve the effectiveness of GPS to determine a more accurate location or change of location of a vehicle. This combined with an ultra precise on-board map, is used to determine where a vehicle is on a street and warn riders if they’re interpreted to be on a sidewalk.
Location-based technology works well for static environments, provided you deploy the additional base stations required, but cities are anything but static. Imagine using GPS instead of your own view of your surroundings to get to a destination within your city. While GPS gives you the location information you need to get where you're going, there are lots of other factors along your journey that need your attention. Pedestrians crossing the street, the speed and flow of surrounding traffic, obstacles in the roadway, these are all factors that are not accounted for by GPS alone. This is why using any combination of technology that doesn’t include a camera-based solution will not provide enough information to effectively understand the surrounding environment and account for the safety of riders and pedestrians.
This type of technology relies on a very precise and static on-board map. Any new bike racks, bike lanes, construction detours, etc., will need to be manually added to the static map and then uploaded to all vehicles. As you can imagine, there is a lot of room here for error. The effectiveness of this technology will likely degrade over time as more and more changes and additions are made to city infrastructure. This is especially true if the vendor doesn’t employ an in-house, local workforce that can quickly and accurately upload these changes, keeping pace with the rate at which they’re happening. It isn’t easily scalable.
The final tool used in sidewalk-riding detection technology is accelerometer-based, which uses the device’s on-board accelerometer to analyze vibrations and inertial measurements to determine a rider’s behavior. This tool has been used with an algorithm that measures the seams, or gaps in the slabs of concrete, used to construct sidewalks to “feel” when a rider is on a sidewalk. Feeling alone, without the use of sight, will only take this solution so far.
The problem here is that not all sidewalks are created equal and cracks or damage to a sidewalk could easily confuse the algorithm. Modern infrastructure looks different in many cases and may not have the features needed to help the device detect a sidewalk. Even when paired with a location-based solution, this technology typically doesn't gather enough information to understand it’s environment to the full potential possible today. The possibility of this technology improving with time is slim to none. It simply doesn’t gather enough data to be used in new ways as algorithms improve.