The MIT Autonomous Bicycle Project

The MIT Autonomous Bicycle aims to bring the convenience of on-demand mobility to bicycle-sharing, allowing more people to enjoy their cities in a new and environmentally friendly way.  In addition, it would serve as an efficient solution for the first- or last-mile,  improving the access to mass transit, and reducing car-dependency for longer trips.   Being an on-demand mobility solution,  users would be able to call for a bicycle trough an app and a bicycle will drive autonomously to wherever they are. Then, it would be ridden just like a regular bike and once the trip is complete, it would go back to autonomous mode either to pick up its next user or to a charging station.

Why bicycles?

According to the latest report from the United Nations, by 2018 cities represented 55% of the global population with this number growing to 68% by 2050. As the world continues to urbanize, cities will have to face new challenges that will require innovative solutions regarding housing, infrastructure, services, and mobility. New mobility demands will need efficient, ecological, inexpensive, and reliable means of transport that ensure a convenient flow of people and goods around cities, minimizing mobility-related problems such as congestion, air pollution, or long daily commute times.

At the City Science group, we envision the cities of the future to be "human-centric", being composed of dense and diverse districts where people can walk or bike from home to school, work, or even shopping. In this new car-less scenario, shared and autonomous micro-mobility systems like the MIT Autonomous Bicycle would work either as a mobility option for short commuting distances or as first- or last-mile solutions in connection to mass transit.

System design and initial prototype

One of the key aspects in the mechanics of making a bicycle drive autonomously is self-balance. To solve this issue, we have designed an innovative mechanism that provides two different configurations: When in use,  the bicycle configuration allows the experience to be the same as riding a  regular bike. Then, for autonomous driving, it transforms into a tricycle that has the necessary stability to drive by itself.   This transition is generated by two linear actuators that dynamically separate and rejoin the two rear wheels as needed.

In the MIT Media Lab City Science group, we have designed, built, and tested a full-scale functional prototype of this mechanism. The prototype also includes the motors for its movement in autonomous mode: a drive motor for the propulsion and another motor for steering.  The current prototype is operated through a remote control; next steps include further integrating the autonomy hardware and software.

Potential benefits

The MIT Autonomous Bicycle can offer an improved user experience by bringing the convenience of mobility-on-demand systems to bicycle-sharing, while solving, at the same time, some of the challenges found in current systems such as the rebalancing problem or the over-quantification of fleets.  

Fleet rebalancing is one of the major issues in today's bicycle-sharing systems. Due to similarities in commuting patterns  - e.g. people moving inbound in rush hour - and other effects such as the presence of hills, bicycles concentrate in some areas of the city at certain times of the day, while other areas remain bikeless. In these situations, finding available bikes or docks becomes an issue, increasing travel time, and causing users’ frustration and loss of reliability in the system.  To mitigate this problem, operators have to redistribute bicycles though the city in specifically designed vans or trucks, leading to a high economic and ecological cost. 

On the other hand,  dockless systems, not being restricted by the number and size of stations, may suffer from over-quantification.  These fleet sizes may exceed the infrastructure capacities of the cities and user demand, causing urban problems  and the flexibility in the drop off location can mean that people will leave bikes piled in the streets and sidewalks. Because of these issues,  some cities have started to limit dockless fleet sizes, while others have forced operators to cease their programs.

With a fleet of autonomous bicycles like the MIT Autonomous Bicycle, the rebalancing problem is lessened as  bicycles would drive by themselves to the next user (or to the region where demand is predicted to occur).  In addition, being more efficient than current systems,  fewer bikes would be needed to meet the same demand. 

For users, to have a door-to-door service without having to deal with the problem of finding available bicycles or docks would greatly improve the bike-sharing experience. Reduced travel times and increased convenience would incentivize more people to use shared-bicycles for commuting, supporting this way a transition towards more livable and sustainable cities.