For a glimpse of transportation systems of the future, look to the trajectory of fleet management tools. Technologies used today for managing fleets of cars, trucks and even bicycles increasingly represent a key to new business models and an essential component of public transit projects now attracting government attention and investment. The sensors, GPS, advanced routing applications and other tools in these management systems are forming the foundation for a new generation of providers of mobility as a service — beyond the old iterations of buses, trains and taxis.
Ryan Chin, a PhD candidate in the Smart Cities research group at MIT, is working in collaboration with General Motors (s GM) and other MIT researchers to develop a stackable subcompact electric vehicle for urban car sharing. He sees transportation increasingly moving towards the service model as a way to reduce reliance on personal vehicles — helping to address urban traffic congestion, encourage adoption of alternative transit and slash greenhouse gas emissions from the transportation sector. But he says “there’s a lot of difficulty convincing automotive companies to adopt this model. They’re product companies,” he said. “That means there’s a big, huge opportunity for other players to come in.”
Mobility on Demand
Chin’s team has been working with a concept called Mobility on Demand, or MoD — a comprehensive system (illustrated at left) in which city residents would be able to rent an electric car, scooter or bicycle when and where they need it in order to bridge the “last mile” gap in many public transit systems (e.g. getting between the subway station and your final destination).
It might sound like some of the other mobility-as-a-service offerings we’ve written about on GigaOM Pro — the sharing services from companies like Zipcar and San Francisco’s City Car Share for example, which have taken off in recent years. But Chin noted an important distinction. While Zipcar users pick up and drop off a vehicle at the same location (two-way sharing), MoD users would be able to pick up a bike or car and then drop it off at any other station in the operator’s network (one-way sharing).
Pieces of the MoD system envisioned in the lab have already made their way into real-world practice — in a bike sharing program in Paris, for example, and a car sharing program in Singapore that lasted from 2000-2006. Daimler (s DAI) has a pilot program called car2go in Germany and Austin, Texas, in which registered members can rent a Smart Fortwo car by the minute, hour or day, and then return it to any unoccupied parking space within a set operation area. But we have yet to see a company do for MoD what Zipcar, U-Haul, Hertz and other companies are now doing for 2-way car sharing: build a lasting business out of it, and push it toward the mainstream.
Several MoD-type services at this point are run by advertisers through public-private partnerships, explained Chin, with companies to operate the system in exchange for advertising space, and city governments providing land in exchange for the transportation service and a potential solution for traffic congestion.
The opportunity presented by MoD tech also extends beyond green transit networks for consumers. Electric vehicle charging infrastructure startup Better Place plans to route subscribers to different charging stations based on factors incuding location, batteries’ state of charge and how crowded a given station is. And a number of apps for smartphones are already starting to take advantage of real-time location-based data about the supply and demand for mobility to help people carpool. Simply put, these services identify and link passengers and drivers with empty seats that are nearby and planning to go in the same direction.
Corporate fleet managers could also see benefits from advances in these types of systems if they offer a way to serve more employees with fewer cars, and thus less cost. Like utilities facing a flood of data from the smart grid in coming years, fleet managers have access to more onboard vehicle data now than ever before, and the mountain of information is set to grow in coming years as telematics units (wireless communication, vehicle monitoring systems and location devices) become more prevalent and electric vehicles are introduced (see our report on IT and Networking Issues for the Electric Vehicle Market). In the wake of all of this information, and the need to manage it, Diego Borrego, founder of fleet management tech firm Networkfleet told Xconomy recently, “Extracting….data from the vehicle, and somehow transmitting it (in conjunction with GPS) and turning it into information that can be used by fleet vehicle managers is the innovation.”
That’s part of how Zipcar has gotten its foot in the door of corporate fleet management. In Washington, D.C., where Zipcar has piloted its FastFleet reservation and management system, the city pays Zipcar a one-time fee of $1,200 a car to install a monthly fee of $115 per car for “black box” devices (a custom circuit board, processor and modem) on its vehicles and then $115 per vehicle each subsequent month to maintain them. Savings — as much as $300,000 in the first six months — offset the new costs. And Zipcar CTO Luke Schneider told us earlier this year that, as the company gets more fleet customers, it will be able to compare the data collected and “sanitized” with the FastFleet system to identify and adopt best practices in similar organizations.
The theoretical ideal for an MoD system, which could reduce reliance on personal vehicles and boost adoption of cleaner transit options, said Chin, is to have a parking spot or space on a bike rack “magically open up as soon as you enter the station,” and have another user ready to check the car or bike out again as soon as it becomes available.
Of course, the theoretical ideal doesn’t happen very often in practice. “So you have to have a buffer, which costs money,” Chin said; an operator is not earning money when the vehicles aren’t in use. Inefficiency comes into the system primarily as a result of asymmetric trips, with an influx of people trying to go to commercial centers from residential areas around the same time. Operators can deploy sensor technology and collect real-time data on their fleet, but that doesn’t solve the problem. “You can sense all you want, but the system goes out of whack,” said Chin.
In Paris, the solution has been to redistribute the bicycles with trucks, while in Singapore employees took the subway to over-supplied stations and drove unused cars to areas with higher demand to “zero out the system” each night, said Chin. Then in addition to managing users and a fleet of shared cars or bikes, the operator also has to manage the redistribution of staff and trucks. “That’s not sustainable,” said Chin, and it lessens the impact of these services as a way to clean up transportation. “What you want is users moving vehicles around, not trucks.”
Need for Better Algorithms
The challenge for modeling supply and demand of the vehicles in these systems, and how variables like dynamic pricing can affect them (similar to the way U-Haul charges more for truck and trailer rentals where departures outnumber arrivals, and less in cities where it’s the other way around), is at least twofold: MoD systems change the way people get around, and they’re so nascent that there isn’t much data about how. More accurate models would use historic transportation data, available from the National Highway Traffic Safety Administration, as well as origin and destination data from an MoD system after it’s been implemented, said Chin. The bike-sharing programs that are starting to crop up in Europe, Washington D.C. and elsewhere provide valuable data for more complex systems. (Cars complicate matters, said Chin, partly because you’re dealing with longer travel distances).
The biggest winner in MoD systems will be the operator with the highest level of service, said Chin: always having a vehicle available within a reasonable time, using the least number of vehicles for the largest number of users. “You’ve basically solved the congestion problem, and minimized land use as well.”
The key to that will be a better algorithm, according to Chin, for mitigating the redistribution problem. “It’s the same thing Google does,” but instead of searching for resources online, the MoD engine would crawl transit-use data. Google won search by building a better engine than anyone else, said Chin. The company that outperforms competitors in MoD, said Chin, will be the one that “builds a better engine based on historic and current data than anyone else.”
Josie Garthwaite is a Staff Writer for Earth2Tech.