IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Data-Driven Research Aims to Solve First/Last-Mile Problem

Together with the city of Atlanta and Georgia Tech, the Socially Aware Mobility Lab uses data and machine learning to look at how on-demand multimodal transit could improve traffic congestion and mobility inequalities.

metrolab img 22
MetroLab Network has partnered with Government Technology to bring its readers a segment called the MetroLab Innovation of the Month Series, which highlights impactful tech, data and innovation projects underway between cities and universities. If you’d like to learn more or contact the project leads, please contact MetroLab at for more information.

In this month’s installment of the Innovation of the Month series, we explore the work of Georgia Tech and the city of Atlanta on the Socially Aware Mobility Lab (SAM).

MetroLab’s Ben Levine spoke with Pascal Van Hentenryck, an A. Russell Chandler III chair and professor in the Georgia Tech Milton Stewart School of Industrial and Systems Engineering and head of the SAM Lab; Jacob Tzegaegbe, senior transportation policy advisor for Atlanta Mayor Keisha Lance Bottoms and part of the SAM advisory board; and Debra Lam, managing director of smart cities and inclusive innovation at Georgia Tech.

Ben Levine: Can you please describe the origin and objective of the Socially Aware Mobility Lab, and who has been involved in the project?

Pascal Van Hentenryck: The Socially Aware Mobility Lab (SAM) aims to transform mobility for all segments of the population and decrease inequalities in mobility while addressing issues such as congestion and sustainability, including reducing greenhouse gas emissions. It focuses on providing new mobility systems to fundamentally improve accessibility to jobs, health care, healthy food and education. It is organized around the concept of on-demand multimodal transit systems that combine on-demand mobility services that serve low-density regions and address the first/last-mile problem with high-occupancy vehicles, like buses or trains, traveling along high-density corridors. The lab has deployed successful pilots in smaller cities and is now working on mobility in larger cities. It is supported by a large NSF Leap HI (Lead Engineering for America’s Prosperity, Health, and Infrastructure) grant to improve American health and prosperity.

Jacob Tzegaegbe: SAM is not a typical research project but part of a larger research-driven partnership Georgia Tech has with the community through the Smart Cities and Inclusive Innovation Initiative. It focuses on the community issue and looks to contribute multidisciplinary, innovative research and development. SAM has established an external advisory board that provides both insights on the project and willing partners for research implementation. External partners include the city of Atlanta, the Metropolitan Atlanta Rapid Transit Authority (MARTA), Gwinnett County, the Atlanta Regional Commission, the Georgia Department of Transportation, the Midtown Alliance, the Atlanta-Region Transit Link Authority, and the State Road and Tollway Authority.

Levine: What are some of the initial outcomes of this project? Have any been surprising or unexpected?

Debra Lam: SAM was able to identify, clean and aggregate many types of mobility and community data. Preliminary results show that on-demand multi-modal transit systems can improve convenience and decrease costs significantly in Atlanta. Moreover, new community ride-sharing systems could reduce congestion by a significant amount using real-time optimization and machine learning. The project also shows that e-scooters may play a fundamentally positive role in the city.

Levine: Can you tell us a bit about on-demand multi-modal transit systems (ODMTS) and how they relate to the SAM Lab?

Tzegaegbe: ODMTS is a novel concept in mobility combining the best of transit and ride-sharing systems. At a high level, it can be viewed as a way to bring transit to the 21st century, using mobile apps and data analytics to design mobility systems that pick up riders at their origins, drop them off at their destinations, and use high-occupancy vehicles to address congestion and economies of scale. They significantly improve convenience for riders, are priced like a regular transit system and are completely synchronized through mobile applications. The city of Atlanta is ultimately interested in improving mobility for our residents and preparing for coming trends, like autonomous vehicles and electrification.

At SAM, we believe that the future of transit blends customer experience with data and technology, like high-performance computing, data mining, machine learning and optimization. ODMTS requires exploration of compelling research questions and provides a direct link to impact people and communities.

Levine: In what ways are on-demand multi-modal transit systems data-driven? What kinds of data are you leveraging?

Lam: The SAM lab leverages gigabytes of data collected daily to design novel mobility systems. It uses transit data from MARTA and Gwinnett County including Breeze Card transactions, automatic passenger counts and cash transactions. It uses vehicle trajectories from the Atlanta Regional Commission, which are unique in the United States. It uses the road network from OpenStreetMap, trajectories of the Bird e-scooters and aggregated Uber data.


Figure 1: Preliminary design of an on-demand multi-modal transit system for Atlanta.

Levine: Can you walk me through this preliminary ODMTS design? (See Figure 1.) What would a “day in the life” be like for someone using ODMTS?

Van Hentenryck: In this design for ODMTS, we are showing how the existing MARTA system can be augmented with high-frequency bus lines and on-demand shuttles. We call the red-dotted locations on the map “virtual points,” and they allow passengers to gather at single locations at the same time, overcoming the common obstacles to carpooling. The dark blue and orange lines extending from the MARTA map are bus lines, the color indicating the frequency (blue means high frequency). The teal lines represent shuttle services, which greatly extend the reach of the entire transit system.

Here is an example of what a user’s experience with the ODMTS mobile app could look like: After entering the application using their smartphone, a rider would choose an origin (a virtual stop) and destination (another virtual stop), a pickup time, and a number of passengers. The profile of the rider would also specify if they need a vehicle with wheelchair accessibility. Once a rider has entered an origin and a destination, they would see a snapshot of their trip, including the approximate wait and transit times. They could then request the ride. The app would assign a driver, and the rider would see a visualization of the shuttle coming as well as the expected pickup time. The name of the shuttle would be displayed at the top of the screen. When the shuttle arrives, the application would notify the rider, who could now board the shuttle. Once the rider is picked up, the vehicle would change color in the app and the rider could follow the ride visually and in real time until arrival. If the trip is multimodal, the app would indicate which buses, trains or shuttles to board at a transfer point and the trip would proceed with the next leg. All the synchronization can be automatically handled through the app.

Levine: Which parts of this project are human-driven and which are technology-driven? What new or innovative technology are you employing?

Lam: Mobility systems are socio-technical systems: They need to focus on people and provide services that will be widely adopted and fully meet their needs. Hence any design starts with what people do, what they would like to do, and how best to serve these needs in an accessible and affordable way. The technology elucidates these needs, plans the mobility system holistically, and operates in real time, adjusting to the needs dynamically. The secret sauce is a tight integration of artificial intelligence and operations research, with optimization, machine learning and massive mobility data sets being the key ingredients.

Needed bus lanes for suburban Atlanta counties

Figure 2: Need for dedicated bus lanes for Suburban Counties. Compares rush hour traffic to times with no traffic. Green: no difference; yellow: rush hour causes up to 13-minute delay; orange: rush hour causes 13 to 26-minute delay; red: rush hour causes 26 to 40-minute delay.

Levine: What are the next steps for this project? What goals do you have?

Van Hentenryck: SAM is trying to raise funds to deploy an on-demand multimodal transit system for Gwinnett County, including its connections to MARTA and to downtown and Emory University using rapid transit buses. If we are successful, this will mean that the concept can be applied to many suburban communities across Atlanta, changing mobility in fundamental ways. SAM is currently doing the design of these mobility systems. The other main goal is to design a mobility system centered around e-scooters for midtown and downtown Atlanta and their connected neighborhoods. Finally, we are looking at other places that could utilize this research, such as military bases and community improvement districts. There are many scales in the city and community to which this research could be applied, and we are excited for the partnership.

Tzegaegbe: We are excited to be partnering with SAM as we continue to think about the future of mobility in Atlanta. Transportation is central to the lives and experiences of people who live and work in Atlanta and we are focused on using these types of mobility approaches to address a range of priorities: more efficient and cost-effective commutes, better coordinated and integrated services, meeting sustainability goals, leveraging this technology to improve access for underserved communities, and improved quality of life.

Lauren Kinkade is the managing editor for Government Technology magazine. She has a degree in English from the University of California, Berkeley, and more than 15 years’ experience in book and magazine publishing.
Sponsored News
As municipalities emerge from shutdowns, slowdowns and travel restrictions due to the COVID-19 pandemic, the transportation sector’s role in supporting the recovery is becoming more apparent.
By now, you’ve probably already heard plenty about data in intelligent transportation. By their very nature, technologies that support intelligent transportation programs capture massive amounts of data.