Making Microtransit Work When Maps Don’t
December 8, 2020 | Company Blog
On-demand microtransit, where shared rides can be booked and drivers respond to requests in real-time, has been a challenge for mobility providers. Balancing supply and demand, ensuring predictable ride and wait times, and enabling operators to maximize revenue with strong fleet utilization are not trivial efforts. Vehicles and drivers need to be routed efficiently and every kilometer without a passenger loses money.
Bestmile’s Fleet Orchestration Platform is designed to meet this challenge. With automated dispatching, ride matching, pooling and routing based on real-time, prebooked and forecasted demand, the platform enables operators to continually optimize services while controlling passenger convenience and vehicle/driver utilization thresholds.
An additional challenge in making dynamic demand-responsive transportation work has to do with mapping and routing technology. Advanced constraint–based routing and customizable maps are needed—functionalities that go beyond what commercial general-purpose routing/navigation products can provide.
Microtransit services and the buses and vans that they typically use face routing constraints and requirements that private autos don’t face. Examples include:
This means that if a public microtransit service relied on off-the-shelf routing and navigation products, they would likely give drivers incorrect routes and turn-by-turn directions.
Fortunately, the Bestmile platform makes it possible to specify the required routing behavior in a very fine-grained, vehicle-specific way, thereby enabling operators to adapt them to each service area and vehicle related restrictions.
Let’s take the FlexiBus service in Switzerland as an example. Transport publics de la region lausannoise (tl), the public transportation agency for Lausanne, piloted an on-demand, station-based, pooled microtransit service with 47 stops in a municipality in the West of Lausanne and wanted to guarantee wait times of no more than 10 minutes.
tl used the Bestmile Fleet Orchestration Platform to orchestrate the service and also used Bestmile’s Traveler App to enable passengers to book rides, view vehicle locations and ETAs, and get directions to stops; and Bestmile’s Driver App to send missions to drivers and to guide them with turn-by-turn navigation.
For the FlexiBus project, three capabilities of the platform came into play.
The first step consists in importing a detailed base map and making on top of it service-specific changes. This includes correcting errors and the ability to make fast updates for temporary changes like construction detours during the project. This global map is the common foundation for all vehicles.
The second step is the definition of an Operational Design Domain (ODD) for each vehicle. ODD is a term used in the field of autonomous vehicles and defines where vehicles can go, how vehicles can move spatially (possible maneuvers at intersections), and temporal dynamics. This includes removing the streets that cannot be used and adding additional routing possibilities such as enabling vehicles to use public transport lanes in case of a microtransit service.
For the FlexiBus project, ODD specifications were also used to define where certain actions like u-turns and left-hand turns could and couldn’t be made, depending on the vehicle, to ensure safety, directional accuracy and route efficiency. When passengers cannot cross the street to board a vehicle stopped on the other side, it is essential to route vehicles such that they arrive at the desired pickup or dropoff location on the right side of the street. To account for this, Bestmile’s routing technology ensures that the vehicles approaches the location from the correct direction.
When a vehicle is leaving a stop, the routing also takes into account the current heading of the vehicle in order to avoid imposing inefficient or prohibited U-turns on the driver.
The third step consists in enhancing time estimates, taking into account all ODD constraints and the vehicle specificities like maximum speed. This enabled tl to project wait times and ride times based on the speed of a bus (vs. a car), also taking into account routing limitations based on vehicle-specific capabilities. The platform also links to third-party traffic information for more accurate estimates.
The result is a routing technology that always give the most appropriate and most efficient directions based on all of the local specificities.
Map editing, ODD-based routing, and accurate time estimates based on local vehicle-specific constraints are not possible when relying on off-the-shelf mapping and navigation solutions.
In the case of the tl FlexiBus project, the service made thousands of on-demand trips with just two vehicles while meeting the 10-minute maximum wait time requirements. Over the course of the six-week project, nearly 3,000 passengers traveled some 7,000 km.
Every city is different, and every mobility service is different. Generic, off-the-shelf solutions–like the most common commercial routing and navigation products in this case—will not always be able to account for all local restrictions and requirements. The ability to adjust services to vehicle-specific constraints will also be important when autonomous vehicles are allowed on public streets, as they will likely bring additional maneuvering, accessibility, and speed capabilities and limitations. Car-like robotaxis, for example, will have different constraints than larger shuttles and buses, and operators will need to be able to design services accordingly.
Fortunately for mobility providers, the Bestmile Fleet Orchestration Platform was built for AVs first, prepared to direct vehicles without a human driver and accounting for the myriad of vehicle and service area constraints that must be considered. The platform has, built-in, the power and flexibility to adapt to permanent and temporary routing and navigation limitations while enabling operators to control passenger and operator performance requirements. These capabilities enabled tl to quickly and easily design and test an efficient and accurate on-demand microtransit service for a specific neighborhood based on local conditions and vehicle capabilities that met promised service levels.