Birmingham, Alabama, United States
Mobility on Demand (MOD) Sandbox Demonstration: DART First and Last Mile Solution
Summary Information
As a part of the Federal Transit Administration (FTA) Mobility on Demand (MOD) Sandbox Demonstration, this pilot study aimed to improve first mile/last mile (FMLM) access to Dallas Area Rapid Transit (DART) transit for all people including individuals with disabilities, improve the experience of transit, increase transit ridership, and provide more transit options for riders in underserved areas. The GoPass application (app), DART’s existing regional mobile app, was modified and enhanced through this pilot to facilitate first and last mile connections to offer transit riders travel options based on price, wait time, travel time, and payment. This study implemented a three-phase approach by leveraging the Application Programming Interfaces (API) of key mobility partners and providers during 2017-2018 in the Legacy West, Plano and Far North Plano (FNP) areas: first, a Microtransit app, called GoLink, was developed; second, a smart app switch which included metadata information connected GoLink with GoPool (a pre-existing carpooling app) was created; and finally, the complete integration of GoLink was accomplished. DART’s GoLink app offered microtransit services through collaboration with a microtransit provider and Transportation Network Company (TNC) to pilot areas that did not have bus services before, and provided mid-day service to restaurants and shops that previously had not been available. Moreover, in Phase 3, the upgraded GoPass app provided on demand travel information for multimodal trips and provided a remittance mechanism for payment. DART also started an UberPool program in three Plano GoLink zones in March 2019 to offer another MOD option at a lower rate (DART subsidized the difference between the DART rate and the actual cost of the Uber trip). Both quantitative and qualitative evaluations were conducted using trip activity data and a user survey to assess performance.
Lessons Learned
- Test drones’ vehicle detection limitations prior to deployment. Results from this study pointed out that the bounding box of the vehicle must have the size of 100 pixels at least, and its smallest dimension must not be less than 8 pixels for accurate vehicle detection.
- Deploy drones prior to actual event requiring traffic management for better validation of vehicle detection and calibration of the traffic stream models. For locations that are pre-selected for drone array deployment, it would be advisable to deploy the drones prior to the actual event requiring traffic management to validate the detection and calibrate the traffic stream models if they will be used for traffic state assessment.
- Use a Graphics Processing Unit (GPU) to accelerate vehicle detection. In this study, increasing the number of processing cores led to less detection time. Therefore, GPU was used to accelerate detection since GPU has a higher number of smaller cores and it can carry out parallel processes more efficiently.
