Fleet orchestration: keys to successful AV mobility service
Anne Mellano, Bestmile Cofounder and VP of Operations, looks at three factors that need to be considered when rolling out autonomous mobility services in ways that will provide the maximum benefit for operators, cities, and travelers.
Much of the attention paid to autonomous mobility has focused on the vehicles: Do they work? Are they safe? When will they arrive? What will be the impact?
What gets less attention is how these vehicles will work together to deliver services that will a) make traffic and air quality better and not worse; b) support rather than compete with existing modes of transit; and c) make all transportation more efficient.
Let’s look at three factors that need to be considered when rolling out autonomous mobility services in ways that will provide the maximum benefit for operators, cities, and travelers.
1: Orchestrate, orchestrate, orchestrate.
For new mobility services to be efficient, autonomous vehicles need to be able to work together as fleets. Like musical instruments, each playing randomly can result in chaos. Look at what has taken place with peer-to-peer ridehailing services. Traffic and pollution are worse as these services have added 5.7 billion miles driven on the streets in the U.S. alone, and each vehicle travels 2.8 miles for every mile it carries a passenger.
Fleet orchestration is more than fleet management. It’s not just telematics and preventative maintenance. It is getting the right vehicles to the right travelers at the right time, every time. For example, if there are four vehicles near a person requesting a ride, how is it determined which vehicle will be assigned to the traveler? The closest? Well, the closest vehicle to the traveler may not offer the most efficient overall experience.
It gets complicated. Determining the best option will depend on the directions the vehicles are traveling, the traffic, the side of the road the traveler is on, the capacity of the vehicle, the amount of battery power remaining in the vehicle, and more. It’s also a matter of distributing vehicles and services so that they aren’t all looking for the same riders at the same times. These are complex calculations that, done properly, can help drive intelligent vehicle allocation that can anticipate and meet demand.
ANNE MELLANO is scheduled to continue the dialog at our Fall event AVT Connect.
View Sessions and Speakers.
2: Services have to be shared to be effective.
Replacing individual vehicles with individual autonomous vehicles might reduce pollution if the vehicles are electric, but will do little to ease congestion. In fact, some think that congestion could get worse as people won’t mind traveling for hours in comfortable auto-lounges, or if it is less expensive for vehicles to cruise than park. Services will need to be shared. But for sharing to work, it has to be more convenient and less expensive than driving. This is where the real value of orchestration comes to the fore.
Making shared services more efficient than driving is even more complex than the example given above. With hundreds of requests for rides, determining how to best pool travelers will require calculating multiple ride times, pick-up times, destinations, and any thresholds set for minimal excess overall ride time against any vehicle utilization requirements. Also, the location of the next ride has to be considered as the cost of any empty kilometers to the next customer will need to be factored into the overall cost. Balancing operator and passenger requirements are critical.
The results of fleet orchestration can be extraordinary. MIT researchers found that 3000 four-passenger cars could serve 98% of the demand otherwise met by 14,000 taxis in New York City, with an average wait-time of only 2.7 min. The study also found that 95% of demand could be met by 2000 10-person vehicles.
To test this theory further, Bestmile data scientists studied taxi data from the city of Chicago and used the Bestmile platform’s ride matching and dispatching algorithms to model how a shared service could perform compared to the taxi service. The study found that 200 six-passenger shared vehicles could do the work of the city’s entire 2700-taxi fleet with an average excess ride time of 6 min and a wait time of 5 min.
3: Fleets need to work with public transit and cities, not against them.
Where ridehailing services have worked independent of transit operators, public transit utilization has decreased, as it has across the board in the U.S. (down 2.9% in 2017). A first-of-its-kind University of California Davis study of ridehailing usage found that 49-61% of passengers used the service in place of public transit.
This is not sustainable. Using autonomous vehicles for shared first-mile/last-mile services, getting people to and from public transport hubs is one of the most exciting use cases for AV technology. Conveniently getting from the train station to work is one of the biggest barriers to public transit use. Hailing an autonomous ride that will get you from the train to the office in a jiffy would likely boost ridership.
Getting it right will require working with cities. A McKinsey report found that Smart City technology has the potential to improve urban quality of life indicators by 10-30%. These are “numbers that translate into lives saved, reduced crime, shorter commutes, a lower health burden, and carbon emissions averted,” the report said. The contribution of mobility services to these figures are in the shorter commutes and lower carbon emissions. But the aggregated value will be realized when all of the technologies that make cities smarter can work together.
The ability for anyone, regardless of age, income, and location, to access affordable mobility is fundamental to the quality-of-life benefits we all seek. But the private sector must work with the public sector to ensure that new services benefit all residents. If low-income residents or neighborhoods are left out of the picture, the cities of the future many imagine will not be realized.
It’s not just about the car
Vehicles alone are not services. There is much to be considered when configuring mobility services at the kind of scale that is being envisioned. We have to raise our vision above the individual vehicles and their exciting potential and think through how fleets of shared vehicles can work together, and how these fleets will work with public transit and other Smart City infrastructure.