Principal Investigator

Sean Qian

Status

Completed

Start Date

Jan. 1, 2016

End Date

Dec. 31, 2017

Project Type

Research Advanced

Grant Program

MAP-21 TSET National (2013 - 2018)

Grant Cycle

2016 TSET UTC

Visibility

Public

Abstract

Mobility Data Analytics Center aims at developing a centralized data engine supported by a web application to manage and analyze multi-jurisdictional multi-modal data for safety, mobility and sustainability, using the City of Pittsburgh and the City of Philadelphia as case studies. This research is a continuation of UTC funded research ‘Mobility Data Analytics Center’ in the years of 2014 and 2015, where we have built models and prototype tools to analyze various sources of data for public transit, parking, and roadway. In this continuation for the next two years, we set our focus on the following four aspects. First, maintenance and enhancement of the data engine and web application. We continue to collect and archive up-to-date data from various data providers in the Pittsburgh region and enhance the web application. Second, we establish a regional multi-modal trip planner (MUMTIP) incorporating a regional truck adaptive routing and scheduling system (TARS). Prediction of travel time and crash risk is central to MUMTIP. Third, we use I-79 corridor from Pittsburgh to Morgantown as a case study to illustrate the mobility impact of autonomous vehicle adaptation. In particular, we consider two possible highway settings (designated vehicle lanes with less required lane width, and existing highway setting) and various autonomous vehicle penetration rates. Fourth, we will build a sophisticated multi-modal transportation network model (MUMNET) for the Pittsburgh region that describes individual travel activities on roadway systems, transit systems and parking systems. The multi-modal network model is the key to systematic planning and operations of transportation infrastructure.    

Description

Mobility Data Analytics Center aims at building a centralized data engine to efficiently manipulate large-scale data for smart decision making. Integrating and learning the massive data are the key to the data engine. The ultimate goal of understanding massive data is to accurately estimate the historical usage of the transportation infrastructure and to forecast its future performance. To efficiently balance the infrastructure supply and demand, optimal decisions on management strategies, policies and adoption of technologies can be made. 

Through MAC we propose to develop a centralized data engine supported by a web application to manage and analyze multi-jurisdictional multi-modal data for safety, mobility and sustainability, using the City of Pittsburgh and the City of Philadelphia as case studies. In the centralized data engine, massive data are stored and managed, and can be further translated into useful information for people who need it: legislators, transportation planners, engineers, researchers, travelers, and companies. Unlike the traditional single computer stand-alone software or tools for data preparation and system design, the data engine relies on web-based data sharing and browser-based human-computer interaction for it to be accessed by users. The web application visualizing data and recommending decisions serves the front end of the data engine. 

This research is a continuation of UTC funded research ‘Mobility Data Analytics Center’ in the years of 2014 and 2015. In the past two years, we have started building the data engine and a prototype web application to demonstrate the feasibility of Mobility Data Analytics Center. We started from the Pittsburgh region where we have close partnerships with many local entities, and have successfully applied our data analytics tools in several case studies.  The main accomplishments are summarized as follows, 

1. We have gained access to various data, which include, but are not limited to, socio-demographic data, traffic speed probe data, traffic incidents, parking meters, transit APC-AVL, weather, social media data. Those data are integrated and visualized through the prototype web application shown in Figure 1 (see
the attached page). 
2. We have analyzed large-scale crash data in PA, and developed an online tool to visualize and forecast crash types, frequencies and severity for each PennDOT-owned road segment. 
3. We analyzed large-scale APC-AVL transit data to provide both travelers and transit agencies fine-grained customizable information regarding transit service performance (efficiency, reliability and quality), and have developed a Transit Service Performance Information and Optimization system (TranSEPIO). Port Authority is the potential deployment partner. This has lead to a PITA project. 
4. We have developed a tool to analyze large-scale parking meter transactions data and to provide estimation and forecast of time-varying parking occupancies for each block of on-street and off-street public parking. Pittsburgh Parking authority, MeterFeeder Inc. and several neighborhoods are potential partners. This has yielded a NSF three-year project regarding parking optimization. 
5. We have built a sophisticated transportation network model for the City of Pittsburgh that describes individual travel activities on transportation systems. Operational strategies and policies can therefore be examined in the network model in terms of system delay, crash risk, reliability, vehicle-miles traveled (VMT), fuel consumption and emissions. It has stimulated a research project with PennDOT (Philadelphia traffic impact study) and one with the City of Pittsburgh (traffic impact study of Greenfield Bridge closure). SPC is another potential deployment partner. 

In the years of 2016 and 2017, this effort will continue with the focus on the following four aspects, 

First, maintenance and enhancement of the data engine and web application. We continue to collect and archive up-to-date data from various data providers in the Pittsburgh region and enhance the web application. The data providers include PennDOT, the City of Pittsburgh, Port Authority, Police Department, Twitter, and other private data sources (such as INRIX and TomTom).  The web application the allows travelers and agencies access historical, real-time, and forecasted traffic metrics (such as travel time, delay, incidents, crash rates, etc.) in multi-modal transportation systems (roadway, parking and transit) will be further enhanced in terms of user-friendly design of GUI and efficiency of data processing. The servers hosting the web application will be optimized for load balancing. We will continue to interview various data resource providers in the Pittsburgh region to enhance the quality and quantity of massive data, including governmental agencies, consulting firms and private data providers. 

Second, establishment of a regional multi-modal trip planner (MUMTIP) incorporating a regional truck adaptive routing and scheduling system (TARS). All users can obtain the historical travel time (by car, truck or public transit) of a past time period or the predicted travel time for the next few hours or a future day on selected road segments. System performance metrics, such as Travel Time Index (TTI) and Planning Time Index (PTI), are provided to all users. The travel time historical and real-time data relies on data feeds from three vendors for car/truck (HERE, INRIX, and TomTom) and PACC for transit. The time-varying travel time is predicted by taking into account daily travel demand characteristics and real-time incidents impact. In addition, public agencies can obtain the historical traffic counts and crash counts of a past time period, the predicted traffic counts and crash counts for the next few hours or a future time period on selected road segments. With predicted multi-modal travel time, crash risk and demand characteristics, a trip planner will be built to recommend optimal routes for cars, trucks and transit. The drivers can input arbitrary origin and destination information as well as their departure times. The system will calculate one or several optimal routes for each mode that minimize travel time/cost and crash risk. 

TARS, a critical component of MUMTIP, is specifically designed for trucks in the region. We will work with PennDOT to obtain and integrate infrastructure data on truck restrictions.  Given desirable departure times and a sequence of places to be visited, TARS recommends optimal routes and schedules for trucks by incorporating real-time and historic travel time, crash risk and considering roadway restrictions. Truck parking availability will be considered in TARS while making recommendations on routes. 

Third, autonomous vehicle impact analysis and a case study: Pittsburgh-Morgantown I-79 corridor
With the mixed traffic flow model and regional network model developed under UTC project of 2015, it is then possible to simulate all trips in the Pittsburgh region, and analyze autonomous vehicle impact. The focus is to use I-79 from Pittsburgh to Morgantown as a case study to illustrate the mobility impact of autonomous vehicle adaptation. In particular, we consider two possible highway settings (designated vehicle lanes with less required lane width, and existing highway setting) and various autonomous vehicle penetration rates. Policy indications can be obtained by systematically simulating vehicles in the regional network.  PennDOT is the potential deployment partner. 

Fourth, establishment of a regional multi-modal network model (MUMNET). We will build a sophisticated transportation network model for the Pittsburgh region that describes individual travel activities on roadway systems, transit systems and parking systems. The multi-modal network model is the key to systematic planning and operations of transportation infrastructure. Integrated corridor management (ICM) is one of the applications of the network model. Operational strategies and policies can be fully examined in the network model in terms of system delay, crash risk, reliability, vehicle-miles traveled (VMT), fuel consumption and emissions. This research will propose a framework and methodology for implementing multi-modal dynamic traffic assignment (DTA) in a general corridor network. The multi-modal DTA embeds a nested-logit-based modal split model for three transportation modes, the public transit, solo-drivers and carpoolers. We will reach out to SPC to better understand their needs related to improving their regional planning network model, and have SPC as a potential deployment partner if possible. 

Leverage: PI Qian has been award $40,000 from IBM to conduct smart cities research, which can be used towards MAC development. The ongoing NSF project regarding parking optimization can partially support the efforts of developing MUMTIP and MUMNET for MAC. The PI is also applying for a research grant ‘MOVE 32: regional traveler information systems’ from Benedum Foundation. Upon award, its budgeted $400,000 can be leveraged to advance the development of those models and tools for MAC. Qian is currently working with FHWA for a potential project on stability analysis of network models. The requested project budget is 80,000, part of which is to financially support students to develop network models that are closely related to MUMNET.

Timeline

Year 1:
1 Data integration and web application enhancement (January - December)   
2 MUMTIP and TARS (June - December)
3 Autonomous vehicle impact analysis (May - August)
Year 2:
1 Data integration and web application enhancement (January - December)
2 MUMTIP and TARS (January - June)
4 MUMNET (May - December)
           
One CEE phd student will work on tasks 1 and 2, and the other CEE phd student will work on tasks 3 and 4.

Strategic Description / RD&T

Deployment Plan

We plan to seek both industrial and federal funding for implementation based on the initial development. While we focus on several particular applications (transit and roadway) to demonstrate the method and leverage our resources, the methodology can be broadly applicable and scalable to other cities or regions, and Mobility Data Analytics Center can interact with other urban systems in the long run, such as water/sewer system, energy system, air quality, etc. This generality will attract attentions from various groups interested in smart infrastructure, green design, and environmental policies. Potential funding agencies/collaborators include the Department of Energy, Department of Transportation, Federal Highway Administration, National Science Foundation, National Institute of Standards and Technology, and Environmental Protection Agency. Local government, communities and foundations including the Benedum Foundation are also potential funding sources. The proposed research is closely related to on-going research at Carnegie Mellon in the context of data science, such as air quality studies, climate change, connected vehicles, autonomous vehicles, energy policies, and infrastructure life cycle analysis. Interactions with those groups will have synergistic effects.

Expected Outcomes/Impacts

The desired results are a web application with the following functionalities, ? Share and exchange data from a variety of sources. ? Data can be visualized from the browsers and exported to several standard data formats. ? Travelers and agencies can access historical, real-time, and forecasted traffic metrics (such as crash rate,
travel time, delay, etc.) and plan their trips. ? A large-scale multi-modal transportation network model for Pittsburgh region that supports decision making for public agencies. ? Enhance regional transportation system mobility and safety by data integration/visualization and decision-making support The underlying models for passengers, cars and parking will allow for easy and flexible extensions in the future to other regions with different data sources, and our analysis tool holds great potentials to support decision making for governmental public agencies, including city, DOT, DOE and EPA.

Expected Outputs

TRID

Individuals Involved

Email	Name	Affiliation	Role	Position
xpi@andrew.cmu.edu	Pi, Xidong	CEE	Other	Student - Masters
seanqian@cmu.edu	Qian, Sean	Carnegie Mellon University	PI	Faculty - Research/Systems
shuguany@cmu.edu	Yang, Shuguan	CEE	Other	Student - Masters

Budget

Amount of UTC Funds Awarded

$84923.00

Total Project Budget (from all funding sources)

$84923.00

Documents

Type	Name	Uploaded
Final Report	Qian2_TSETFinalReport.pdf	May 7, 2018, 4:24 a.m.

Match Sources

No match sources!

Partners

No partners!