Autonomy in shared micromobility
18 September, 2019
According to the Godwin law of mobility, ” As a mobility discussion grows longer, the probability of broaching the subject of Autonomy approaches 1″ 😉 . With multiple recent announcements in the micromobility sector, it is getting an even bigger place in the discussions around shared services.
2019, an autonomous year
My first real meeting with autonomous micromobility was in San Francisco last January when I talked with the founders of Weel Autonomy, a Seattle-based start-up aiming to develop a fully-autonomous e-bike. The photos and videos I saw were just incredible…
At the very same time, Uber was creating its “Micromobility Robotics” branch in order to develop autonomous technology for e-bikes and e-scooters, while actors such as Tortoise or Scootbee emerged.
On August 16th, Segway-Ninebot made big waves at its product launch event including the T60 semi-autonomous e-scooter. Its main feature is to use AI in order to avoid obstacles and drive itself back to a charging station. With a remote driver.
Shortly after, Bolt Mobility also released a tweet showing the latest results of its R&D department, an “autonomous scooter doing parallel parking based in server computed path”.
What’s the point?
Ok, all these videos are just cool! But why do we need autonomous technology in micromobility? The current goal of companies involved in its development is to improve the operational costs of running a shared mobility service: charging, balancing/parking, maintenance.
As a major part of the unit economics of a shared e-scooter service goes into charging, autonomous vehicles are a way to address that: a vehicle will drive itself to a charging point (the main argument of Segway Ninebot T60). But it implies the use of specific charging devices: we can identify induction charging plates in the Segway’s video, but we can also imagine that wired charging solutions might emerge. The charging point remains to be designed.
It is very much related to the question of vehicle’s balancing and parking, as the charging point will be parking spots as well. The vehicles should be able to balance the supply/demand by moving to the appropriate areas and park properly, matching the rules that are getting stricter and stricter in most cities.
Finally, autonomous technology will help to gather the vehicles directly in a workshop or to be collected to get there for maintenance. In all cases, the goal is simply to avoid as much as possible human actions and trips to rationalise the costs.
Is it really for tomorrow?
As for cars, we are in between autonomy levels 2 and 3: vehicles are not yet aware of their full environment, and human supervision is still needed.
Autonomy implementation might nonetheless happen soon, as the current challenge is less demanding than in the car industry: the need for micromobility is for trips without riders and happens at low speeds. On the other hand, the autonomous e-scooters or e-bikes will navigate on pedestrian areas which are less ordered than occidental roads, making more difficult the behaviour prediction.
Many legal issues will also have to be answered: who bears the responsibility for anything that could happen when the vehicle is moving on its own? How will the vehicles fit in the urban (lack of) rules towards light/slow autonomous vehicles?
I honestly think that we are still far away from seeing swarms of bikes or scooters moving by themselves in the streets, mainly due to legislation barriers. Before implementing autonomous technology on hardware, easier solutions will be adopted.