New publication on reinforcement learning in route selection

A major issue in the development of reinforcement learning algorithms for autonomous vehicles is the need to make them reflect the preferences of travelers better. One of the ways to do that is to incorporate user schedule preferences under reliability-based route selection criteria into the learning mechanism. This work led by Jinkai Zhou investigates the potential of such an integration. We used data collected from queries from Google Maps to mimic airport shuttle services to train a multi-armed bandit algorithm to see how it is impacted by the consideration of on-time arrival reliability. The work was funded by NSF CMMI-1652735.

https://journals.sagepub.com/doi/full/10.1177/0361198119850457

New research for evaluating Mobility-as-a-Service systems

Mobility-as-a-Service will only work if cities can evaluate markets of multiple operators, and consider not just ridership but also the incentives for transferring costs (fares, access/egress, wait time, in-vehicle time) between operators and travelers. Saeid Rasulkhani and Dr. Chow just published a new modeling framework that explicitly considers these trade-offs, developed with funding from NSF (CMMI-1634973). 

Link to paper: dx.doi.org/10.1016/j.trb.2019.04.008

C2SMART research to be presented at INFORMS TSL Workshop in Vienna

Research from the BUILT lab on “Doubly-constrained rebalancing for one-way electric carsharing systems with capacitated charging stations”, from Ted Pantelidis, Li Li (NYUAD), Tai-Yu Ma (LISER), Joseph Chow, and Saif Jabari (NYUAD) has been accepted for presentation at the INFORMS TSL Workshop in Vienna. The workshop’s theme is ““Transportation in the sharing economy”. The project is funded by C2SMART with data shared by BMW ReachNow.

Preliminary work for this project was previously presented at the 98th Annual Meeting of the TRB in Washington DC.

BUILT Lab to present NSF research at TRISTAN X at Hamilton Island

The work from student researchers Saeid Rasulkhani and Ted Pantelidis will be presented at the Tenth Triennial Symposium on Transportation Analysis (TRISTAN X) on June 17-21 at Hamilton Island near the Great Barrier Reef in Australia. The NSF-funded (CMMI-1634973) research is entitled “A many-to-many stable matching cost allocation model for multimodal Mobility-as-a-Service”, where we develop a novel methodology to extend earlier work to handle cost allocation analysis for multiple operators splitting a traveler’s trip. This ongoing work has resulted in major breakthroughs in facilitating design of integrated services between different operators and transport agencies within a true “Mobility-as-a-Service” setting, providing to them a tool like how the classic “traffic assignment problem” helped roadway planning in the last few decades. We are finalizing our computational experiments and will be submitting this to a journal for publication.

Nick Caros’ work featured in a TEDx talk by NEXT co-founder Tommaso Gecchelin

Nick’s thesis work, which involves simulation of en-route transfers to better understand their impacts under different transit operating strategies, was recently featured in a TEDx TUM talk by Tommaso Gecchelin, a co-founder of NEXT Future Transportation.

Nick Caros was a MS student in the Transportation Planning & Engineering program in the Department of Civil & Urban Engineering. He worked as a research assistant in the BUILT lab through funding from C2SMART and completed his MS degree with a thesis in January 2019. He has one conference proceeding and 2 manuscripts under peer review in international journals from his time here.

BUILT lab to present at ISTTT 23 in Lausanne, Switzerland

Brian Yueshuai He and Prof. Chow’s work on “Optimal privacy control for transport network data sharing” has been accepted for a poster at ISTTT 23, the bi-annual international symposium on transportation and traffic theory. 

The International Symposium on Transportation & Traffic Theory series is since its first issue in 1959 the main gathering for the world’s transportation and traffic theorists, and for those who are interested in gaining (or contributing to) a deeper understanding of the field. The Symposium deals with both scientific and operational aspects of transportation and traffic, spanning all modes of transport, and covering freight as well as private and public transport.
In more 30 podium presentations the attendants will be informed about the latest scientific insights on transportation and traffic theory. For more information about the symposium, see here. 

 

New publication on designing autonomous vehicle car clubs

In this joint work with Prof. Allahviranloo from CCNY, we studied the design of autonomous vehicle fleets for households purchased shared ownership, like stocks, of each vehicle. Under such a system, owners can therefore trade stocks of these time slots, and the impact of a time slot on a user’s activity schedule needs to be accounted for. This work should be of interest to car manufacturers like Ford and GM as they move into these types of AV business models in the future.

The paper can be accessed here: https://www.sciencedirect.com/science/article/pii/S0968090X18316553

Paper review results for 98th TRB Annual Meeting in January 2019

We have received results for our submissions to the TRB Annual Meeting in January 2019: six papers have been accepted for presentation. Congrats to all the researchers involved!

  1. Effects of violent crime and vehicular crashes on active mode choice decisions in New York City — Nick Caros, Joseph Y. J. Chow
  2. Quantifying the effect of cyclist behavior on bicycle crashes and fatalities — Omar Abou Kasm, Ziyi Ma, Joseph Y. J. Chow, Ali Diabat
  3. Multi-armed bandit on-time arrival algorithms for sequential reliable route selection under uncertainty — Jinkai Zhou, Xuebo Lai, Joseph Y. J. Chow
  4. Effects of charging infrastructure and non-electric taxi competition on electric taxi adoption incentives in New York City  — Jaeyoung Jung, Joseph Y. J. Chow
  5. Adapting the business model canvas entrepreneurship tool to assist transportation technology transfer — Shayan Khan, Will Bierds, Jack Bringardner, Joseph Y. J. Chow
  6. Optimal queueing-based rebalancing for one-way electric carsharing systems with stochastic demand — Tai-yu Ma, Ted Pantelidis, Joseph Y. J. Chow