The Philadelphia region has a large and complex freight transportation network that includes more than 1,000 miles of the National Highway System and 9.8 million vehicle-miles of daily truck travel. The mobility of commercial trucks and the efficiency of freight infrastructure are essential to regional transportation infrastructure planning and economic development. Unfortunately, characteristics of freight demand, such as when and how trucks travel, freight destinations and truck routing behavior, are unclear. This becomes the main hurdle for improving truck mobility. There lacks of freight models that predict the mobility of truck demand induced by ‘what-if scenarios’, such as roadway construction, new freight terminals and land-use change. This research will analyze the freight movements from the intermodal facilities in the Delaware Valley Regional Planning Commission (DVRPC) region, including ports, airports, and rail terminals, to understand travel destinations, travel routes, touring behaviors, time of day of travel and other travel patterns of the freight truck movements generated from these facilities, using the truck GPS data purchased and provided by DVRPC. By incorporating the characteristics of truck demand, this research will also develop a regional network model that encapsulates route choices of car and trucks, and estimate/predict high-granular car and truck network flow. As a proof-of-concept experiment, we use this network model to forecast the traffic conditions of trucks induced by the closure of an I-95 highway segment between Ben Franklin Bridge and Broad St. In addition, we develop a prototype web application to integrate and visualize truck demand characteristics, the regional network model and modeling outputs. This tool will be provided to DVRPC for their decision making on freight planning and operation.
Background: The Philadelphia region has a large and complex freight transportation network that includes more than 1,000 miles of the National Highway System and 9.8 million vehicle-miles of daily truck travel. The mobility of commercial trucks and the efficiency of freight infrastructure are essential to regional transportation infrastructure planning and economic development. Unfortunately, characteristics of freight demand, such as when and how trucks travel, freight destinations and truck routing behavior, are unclear. This becomes the main hurdle for improving truck mobility. In addition, how roadway construction projects would impact freight mobility is unknown. For instance, during the reconstruction of the I-95 corridor in the region, accommodation of truck demand, including planning of construction projects and detour strategies for truck demand, is of critical importance to freight mobility, and ultimately to the economic prosperity. Accurate modeling of freight transportation, on both demand and infrastructure, is needed to support those activities. Emerging truck data provides a unique opportunity to learn, estimate and predict truck demand, and furthermore to establish a regional freight transportation model to support regional planning and development.
Delaware Valley Regional Planning Commission’s (DVRPC) existing freight demand model is limited to modeling light and heavy truck trips within the context of a traditional 4-step model. This model is based on a truck survey conducted in 2000 that determines truck trip generation rates and trip length frequency distributions. Vehicle classification traffic counts on several typical weekdays were collected and used to calibrate the model. While the existing model is able to reasonably predict the average truck demand at the level of traffic analysis zones (TAZs), it is unable to encapsulate truck flow in high spatio-temporal granularity. Given sparse (and sometimes synthesized) data, the model is also unable to predict the mobility of truck demand induced by ‘what-if scenarios’, such as roadway construction, new freight terminals and land-use change.
Main modeling challenges:
• Extensive amounts of traffic data for both cars and trucks has become available, but how to integrate them into truck demand modeling and understand truck demand is unclear.
• Traffic congestion is endogenous to truck routing and operations, which in return considerably affects traffic congestion. A challenge lies in modeling of the mixture of truck flow and car flow that can be validated by real-world data.
• Modeling trucks in the DVRPC regional network raise significant challenges to account for high spatio-temporal granularity, while ensuring computational tractability.
Objectives: This research project will analyze freight transportation data and enhance the regional freight transportation model to better estimate and forecast truck travel in the Philadelphia region.
1. Analyze truck demand characteristics: analyze the freight movements from the intermodal facilities in the DVRPC region, including ports, airports, and rail terminals, to understand travel destinations, travel routes, touring behaviors, time of day of travel and other travel patterns of the freight truck movements generated from these facilities, using the truck GPS data purchased and provided by DVRPC.
2. Visualize and understand truck demand: develop a web-based application to visualize the truck demand characteristics.
3. Develop integrated regional network models for cars and trucks in the DVRPC region: with the new truck demand model, this research also plans to develop a regional network model that encapsulates route choices of car and trucks. This model can be used to predict the mobility of truck demand induced by ‘what-if scenarios’, such as roadway construction, new freight terminals and land-use change.
This project primarily works with three data sets: vehicle classification traffic counts that include counts of cars and trucks, INRIX traffic speed data for both cars and trucks at the level of traffic management channels (TMCs), and sample second-by-second GPS trajectories data for trucks. All the three data sets cover the DVRPC region.
Task 1: Integrate various data sources for in-depth freight data analytics. Develop a web application for freight data visualization and descriptive analysis.
Establish a refined GIS model for the DVRPC region. The GIS database include street names, street levels (highway, major arterials, minor streets, alleys, etc.), the number of lanes, and speed limit, integrated with vehicle classification traffic counts at fixed locations, as well as INRIX traffic speed data for road segments of National Highway Systems. Other freight infrastructure data are also integrated into this GIS model, which include all intermodal facilities in the DVRPC region, such as bridges, ports, airports, waterways, freight centers, and rail terminals.
The GIS model will be built into a web application based on OpenStreetMap to visualize those infrastructure networks and their respective descriptive statistics. This web application complements and advances the existing freight mapping tool developed by DVRPC (https://www.dvrpc.org/webmaps/phillyfreightfinder/) by 1) integrating truck movements and demand data into the same GIS platform along with user friendly interfaces; and 2) providing in-depth analytics for high-granular truck travel time and counts data at the 5-min level. Examples include truck travel bottlenecks, reliability, restrictions on weights and under clearance.
Task 2: Analyze and visualize truck GPS trajectories data.
This project will purchase commercial truck GPS trajectory data and analyze freight truck movements originated from intermodal facilities and major freight centers in the DVRPC region. The goal is to better understand the travel and mobility characteristics of freight trucks in term of time of day distribution, travel destinations, travel routes, stop locations, touring behaviors, travel speeds and bottleneck locations, as well as their impacts on the regional transportation network.
The web application will be further expanded to allow the planners and researchers to query, visualize, and analyze the truck GPS trajectories data, as well as characteristics of freight trucks.
Task 3: Modeling of freight truck trip generation at intermodal facilities and the mixture flow of cars and trucks in the regional highway network
We will use a mesoscopic network analysis methodology to conduct this research. CMU Mobility Data Analytics Center uses a dynamic network analysis tool (MAC-POSTS) capable of estimating network-wide traffic impact for any general networks consisting of freeway, arterials and local streets. It has the capacity of modeling dynamic traffic evolution for both trucks and cars, with the consideration of both truck and car routing. It adopts state-of-the-art traffic models and is much more computationally efficient than other microscopic models that are extremely labor intensive to build.
This project will then investigate the feasibility of using the sample GPS data to estimate the freight truck demand in the DVRPC region by integrating it with other freight activities data (e.g. truck counts, achieved truck speed data, and annual commodity flow data). A novel multi-data approach will be developed, as a complement to the conventional travel demand forecasting approach used by DVRPC, for the demand estimation. The sample GPS data are assumed to represent a specific group of truck demand across all TAZs in the region, which can be scaled to true truck demand, bounded by other sample and census freight data. The freight truck demand will be estimated in such a way that the estimated travel time and flow rates of both cars and trucks can approximately match those observations collected in the Tasks 1 and 2.
Through the regional network model, we simulate individual cars and trucks in the reginal network, and model their route choices, travel time and mixed traffic flow conditions. The results include the travel time, travel delay, vehicle-mile-traveled and emissions, and can be presented and estimated at different scales, from each road segment and intersection by time of day, to the entire network of the peak hours.
In addition, the web application will be further advanced to incorporate visualization of car and truck flow in the network, as well as the animation of traffic evolution over time of day. It will also provide interactive user interfaces to visualize model outputs from DVRPC’s regional planning model and MAC-POSTS’ dynamic model to better understand the current truck movements in the DVRPC region.
Task 4: Modeling of the future traffic conditions with the presence of the I-95 closures
As a proof-of-concept experiment, we extend and apply the model calibrated in Task 3 to forecast the future traffic conditions for both cars and trucks induced by the closure of a I-95 highway segment between Ben Franklin Bridge and Broad St near Downtown Philadelphia. The traffic impact can be measured by time-of-day performance metrics at both the street level and the regional level, such as total traffic delay, average travel time, emissions, energy use, vehicle-miles traveled, etc. Those results will again be added to the web application for visualization and animation.
Each task will take 3 months
The proposed research holds great potential to bring decision making methods and social benefits to regions, municipalities, cities, local communities throughout the Commonwealth. When successfully built, we will be actively looking for other industrial partners, engaging other governmental agencies (such as Southwestern Pennsylvania Commissions, PennDOT) in Pennsylvania, and other planning stakeholders in the nation. In addition, we will immediately interview cities and municipalities to debrief our results and models, gauge range of opportunities, and barriers to data sharing and system deployment. We will be actively looking for research funding from those agencies to integrate freight data to existing platforms that focus on passenger cars in the Commonwealth and provide planning tools. In the long term, we plan to build a statewide Freight Mobility models to serve not only individual cities/communities, but also the entire state. We will develop a memorandum among those agencies to ensure their interests and participation in developing this statewide freight modeling platform.
Expected Accomplishments and Metrics
The anticipated results are 1) analyze the freight movements from the intermodal facilities in the DVRPC region, including ports, airports, and rail terminals, to understand travel destinations, travel routes, touring behaviors, time of day of travel and other travel patterns of the freight truck movements generated from these facilities; 2) a web-based application to visualize and understand the truck demand characteristics; and 3) integrated regional network models for cars and trucks in the Delaware Valley Regional Planning Commission (DVRPC) region that is essential for regional planning and decision making.
||CEE & Heinz
||Faculty - Untenured, Tenure Track
Amount of UTC Funds Awarded
Total Project Budget (from all funding sources)
|Data Management Plan
||March 18, 2019, 9:57 p.m.
||March 18, 2019, 9:59 p.m.
||Sept. 23, 2019, 8:27 p.m.
||March 23, 2020, 10:12 p.m.
||July 7, 2020, 12:35 p.m.
||Emerging Mobility Systems
||Dec. 2, 2020, 9:22 a.m.
No match sources!
||Deployment Partner Deployment Partner