This month’s CIVIC Stage 2 Innovation of the Month installment highlights a project called “Piloting On-Demand Multimodal Transit in Atlanta,” which is improving accessibility to jobs, health care and decent groceries for communities that are poorly served by transit and home to low-income populations. The goal is to enhance mobility for all population segments in Atlanta in a convenient, economically sustainable and environmentally friendly way. MetroLab Network’s Elias Gbadamosi spoke with the team’s civic and academic partners about their engagement process and implementation plan in Stage 2 of the Civic Innovation Challenge.
Elias Gbadamosi: Could you tell us briefly what your project is about?
Subhrajit Guhathakurta: The overarching objective of the project is to investigate a potential future for transit systems where train and bus routes are complemented by on-demand dynamic shuttles that act as feeders to and from the fixed network. The resulting On-Demand Multimodal Transit System (ODMTS) addresses the first- and last-mile problem that plagues transit systems across the country. It promises to transform mobility for many population segments, particularly in areas that are currently underserved by transit.
Anthony Thomas: An ODMTS should also reduce the convenience gap between the use of transit and driving a personal car, reducing congestion and greenhouse gas emissions. To evaluate the potential of ODMTS, the project is piloting MARTA Reach in three different zones in Atlanta: Belvedere Park (a mixed-use neighborhood), Fort Gillem (a job center with limited connections to transit) and West Atlanta (a community that is underserved by transit).
Pascal Van Hentenryck: The pilot has the following goals: increasing access and equity in mobility, reaching new transit riders and improving service for existing riders; understanding the needs of riders; and determining the cost effectiveness of an ODMTS.
Gbadamosi: Who are your partners on this project and what specific insights and/or skill sets do they bring?
Guhathakurta: The project is a collaboration between the Metropolitan Atlanta Rapid Transit Authority (MARTA), the Georgia Institute of Technology and the Atlanta Bicycle Coalition (ABC), a transit advocacy group.
Thomas: MARTA is one of the largest transit agencies in the country and serves a complex region that is geographically widespread with many job centers. MARTA brings decades of experience in running transit systems, coordinating various services and stakeholder relationships.
Kari Watkins: The Georgia Tech team is a pioneer in designing and operating ODMTS; they provide the technology to support the pilot, including the mobile applications and the cloud computing services. They also provide expertise in data collection and evaluation, routing and dispatching algorithms, community engagement, and urban planning. The ABC team provides unique insights into the needs of the communities, including flexible payment methods, multi-language applications and more.
Van Hentenryck: This is the first project the team has worked on that is a true partnership between an academic institution and a transit agency. The interactions between Georgia Tech and MARTA are truly synergistic. Georgia Tech lacks the expertise to run a transit service on the street and could not pilot our technology without MARTA. At the same time, MARTA does not have the in-house expertise to develop the routing software. This partnership has been necessary to get the pilot off the ground so quickly, and it will be required to sustain it longer term. The team is also fortunate to have ABC involved to help in the co-development of the service with the public, with a focus on equity and safety. ABC works directly with communities that have not been well-served by the car-dominated transportation system that has evolved in the U.S. Their insight into how to run a people-based transit system has been immensely valuable.
Gbadamosi: What information did you need to gather to effectively design such a large-scale transit system? Were there challenges in gathering that information?
Thomas: MARTA provided ridership data for several years, obtained from Breeze card transactions. This was complemented by travel trajectories provided by the Atlanta Regional Commission, which has built a high-fidelity activity-based model of mobility for the region. The Breeze transaction data captures individual legs, and the team had to develop algorithms to recover the actual origin and destination of each ride.
Van Hentenryck: The design of an ODMTS for the broad Atlanta region is an extremely challenging task, requiring state-of-the art optimization algorithms that use this data as input to produce the network of proposed fixed routes and size the number of shuttles needed for ensuring given performance metrics in terms of costs and convenience. The design is then evaluated by a high-fidelity simulator that uses real-time dispatching and routing algorithms to match riders and shuttles and determine the routes for each trip.
Guhathakurta: The project is particularly interested in determining the value of an ODMTS in a transit-poor community and in connecting people to jobs. It is, however, challenging to anticipate the ridership in underserved communities, and it is a key objective of the project to elucidate the mobility demand in these neighborhoods.
Gbadamosi: One of your goals is for this system to be very flexible to the changing needs of residents. How have you designed flexibility into the system?
Van Hentenryck: Flexibility has been a key concern of the pilot since its inception. MARTA Reach is fully integrated with regular MARTA services. Riders can use their existing payment methods and discounts and pay the same price ($2.50) as regular MARTA trips; they can transfer from MARTA Reach as if the shuttles were a traditional bus/train leg, providing a complete integration of the pilot in MARTA operations.
Thomas: The pilot can relocate vehicles from zone to zone to better serve the demand and new virtual stops can be added to ensure that riders have easy access to the transit service.
Gbadamosi: How have you approached equity and accessibility in your design?
Guhathakurta: Equity is at the forefront of MARTA Reach and has led to many critical decisions including:
- MARTA Reach is offered in multiple zones including a residential, lower-income community of color, an industrial zone, and a mixed-use zone.
- The price of the rides is the same as for the existing transit system and riders can transfer in the traditional way without paying multiple fares.
- All vehicles are wheelchair accessible and can serve two wheelchairs at the same time.
- The mobile applications have been simplified as much as possible to make them accessible to as many riders as possible.
- For individuals who do not have a smartphone, or prefer not to use one, MARTA Reach established a call center that can make trip requests on behalf of riders.
Gbadamosi: Your team hit a huge milestone recently, getting a $1 million National Science Foundation Civic Innovation Challenge award. How has your project evolved since that time, especially in terms of goals and objectives?
Watkins: There is a substantial difference between academic research and a pilot serving real customers and stakeholders. The project has transitioned from research into deployment, application development, community engagement and customer satisfaction. For the first couple of months after the grant was awarded, the team met about five times a week in different working groups (e.g., technology, community engagement, operations).
Guhathakurta: Mobile applications must work in all kinds of edge cases, since riders and drivers always behave unexpectedly. Supervisors must be provided with excellent tools to monitor the system. The cloud computing platform should be available 24/7 and be robust. Community engagement is always more challenging than expected, and the project needs to redouble its efforts in this area. Drivers need to be trained and training material needs to be developed.
Van Hentenryck: The early phase of the pilot also identifies areas where operations can be improved based on demand and driver behavior. It is significantly more work than expected, and also a very different type of work. But it is also exciting to see the difference the pilot makes in the lives of people. Early adopters have been really loyal.
Watkins: The project has also evolved to include major evaluation tasks to ensure that ODTMS is sustainable, scalable and transferable. It is one thing to use NSF funding to run a pilot in one area of MARTA’s service area that was previously not well-served by transit and have it be successful. Showing that it is viable for MARTA to continue to run given their limited funding budget is quite another. The team’s vision for ODMTS is to help MARTA achieve a transit system that is highly efficient and usable by replacing some low ridership fixed-route services with ODMTS, thus scaling up the on-demand portion over time. And finally, the team is collecting data to bring broader lessons learned to the transit industry and transfer the knowledge gained to other agencies that want to try on-demand services.
Gbadamosi: How will residents engage with the end product of your work? What goals do you have for user experience?
Thomas: Patrons will be able to use MARTA Reach as a part of their daily travel to complement existing fixed route services. The pilot will reveal whether MARTA should offer such a service across the entire network. The pilot is also relevant to MARTA’s NextGen bus project — how fixed-route bus service looks in the future must definitely consider how it interfaces with on-demand services.
