HART AV

A project posed by Honolulu Authority of Rapid Transportation (HART) of hypothetical autonomous vehicle (AV) rideshare system in Hawaii.

New transportation modes have been recently emerging within Hawaii from Biki bikeshares to HART’s developing skyline. With talks on a developing autonomous shuttle system at the University of Hawaii at Manoa (UH Manoa), the potential value of an AV rideshare system was to be discussed. This project simulates the AV rideshare from collected data on household activities, and compares the performance of AV simulations to the status quo as well as hypothetical skyline scenarios.

Conducting this project was my first step into the field of research and applying data science and data analysis techniques to other fields. I had learned the difficulty of picking up the disciplines of other fields with no prior knowledge, as before doing this project I had no knowledge of transportation engineering methods and research. Although difficult, my adaptability skills were improved throughout this project as I went from no knowledge on transportation research near the start of the project to writing a full report and presented.

Complex heuristics adapted from vehicle routing problems were implemented in Python. Many simulations using actual data were produced, as well as simulations as if they would benefit from the skyline as if it fully developed. Using regression analysis and hypothesis testing we were able to show that an AV rideshare system with these heuristics would be more efficient than both the status quo and the developed skyline, further pushing the potential for these AV rideshare systems.