Paratransit Service with Automated Dispatch System in South Dakota Led to Over 78 Percent Reduction in Reaction Time for Trip Data Transfer.
Prairie Hills Transit Piloted Artificial Intelligence (AI) Technologies in a Paratransit Dispatch System in the Rural Western City of Spearfish, South Dakota.
Spearfish
Providing a Dynamic, Data Driven Micro Transit Service with Smart Dispatch Using Artificial Intelligence
Summary Information
Prairie Hills Transit (PHT) provides cost-effective, reliable transportation in rural western South Dakota, including a service area that includes nine counties and 15 communities. PHT has a fleet size of 55 vehicles that consist of cutaway buses for wheelchairs and five-point child seats, minivans, one trolley, and transit vans. This project implemented an innovative technology with an artificial intelligence (AI) dispatch to provide a dynamic, automated, and data-driven paratransit service in rural western city of Spearfish, South Dakota, to accomplish the goals of the Federal Transit Administration’s (FTA) Accelerating Innovative Mobility (AIM) initiative. This new service, named DISPATCH360, aimed to potentially optimize the cost of managing transportation operations. The evaluation period of the automated system lasted one year from April 2022 to April 2023, with an additional year-long period of system design from February 2021, prior to going live in April 2022.
METHODOLOGY
The automated dispatch system presented in this study used real-time data from the field, including driver apps, consumer apps, text/voice notification systems. Essentially, the design of the automated dispatch system involved a Global Scheduling Engine (GSE) to utilize all available resources most efficiently, such as the actual street network in the service area, physical barriers, speed parameters, time of day, and appropriate times for the boarding and alighting of passengers. The AI-based dispatch system demonstration was evaluated using quantitative system data, as well as qualitatively with expert interviews and an institutional group discussion with dispatchers and drivers.
FINDINGS
- Results showed that the automated dispatch system led to over 78 percent reduction in reaction time for trip data transfer (transferring a trip into a segment and being visualized by the driver), from 45 seconds to less than 10 seconds, comparing before and after periods.
- Results also indicated a 9.8 percent increase in the number of on-time drop-offs, from 100,887 to 110,736 trips.
- Similarly, a 10.9 percent increase was observed in the number of on-time pickups, from 90,139 to 99,915 trips.
