Skip to main content

BEAM input generation and low exposure routing model

Cite this dataset

Liao, Yejia; Luo, Jill; Hao, Peng; Boriboonsomsin, Kanok (2021). BEAM input generation and low exposure routing model [Dataset]. Dryad.


The Low Exposure Routing (LER) modeling methodology has already been evaluated at the vehicle level in our previous work. In this study, we propose a novel framework to quantify the impact of LER at the transportation system level applying different technology penetration rates. Under the framework, we employ the truck origin-destination from regional transportation demand model, to generate truck trips in the City of Riverside. Then, the calibrated BEAM model, an agent-based simulation, simulates trips through reinforcement learning and dynamic daily planning technique to reach maximum utility at the transportation system level. Finally, it shows that with a 100% penetration rate of LER and a strict 10% time increase threshold, the air pollutant exposure reduced up to 16% at city level with a slight trade-off of travel time.


In this project, we aim to tackle this challenge by developing an activity-based traffic simulation model for the City of Riverside, CA and evaluate the transportation system-level impacts of the truck low-exposure routing (LER) technology for mitigating the impacts of truck emissions on communities. Based on the existing passenger transportation model developed, truck demand and trips are generated using Southern California Association of Governments (SCAG) model. Based on the long-range regional transportation plans in SCAG, we extract zone-based truck trips origin-destination table; thus, we got daily passenger car and truck trips for Riverside City. Since zone-based origin-destination (OD) table offers no exact coordinates and time of a day, this study utilizes employment data provided by the Employment Development Department (EDD) to give deterministic truck trips coordinates and utilizes PeMs data to calibrate the exact time of travel. Those trips are used as input for BEAM, a framework for a series of research studies in sustainable transportation. We then apply activity-based traffic simulation and Low Exposure Routing to Heavy Duty-Diesel Truck (HDDT) trips in the city, showing that the communities will benefit from reduced exposure to air pollutants by adjusting HDDT routes. Finally, we evaluate LER at the transportation system level under different technology penetration rates. 

  1. (hhdt_od_10.csv) Zone-based Origin-Destination table (OD TABLE) from SCAG is necessary for generating three different truck trips: Heavy-Duty Diesel Truck Trips, Medium-Duty Diesel Truck Trips, and Light-Duty Diesel Truck Trips.
  2. (pems_output_i10e.xlsx) PeMs collects real-time traffic data from over 39,000 individual detectors along with the freeway system crossing major cities, monitoring traffic volume by categories, such as number of trucks on the road. Thus, we can estimate deterministic time of travel by reading PeMs data. Since PeMs data is one-hour interval format, we randomly distribute these trips into corresponding hour.  
  3. (truck_related_employment.csv) The Employment Development Department (EDP) record business companies in Riverside City, which we can use them for potential truck trip position generation. Thus, for the truck trips inside Riverside City, we assign coordinates of companies in the corresponding zone to the truck trips as their origin and destination positions. For those truck trips just crossing city, we assign the closest boundaries points coordinates to them.
  4. (riverside_new.xml) The network files for Riverside City is derived from, which is osm (OpenStreetMap) format file. The raw network file is opened in Java OpenStreetMap Editor and transformed into pbf file. After we place this pbf file in the R5 folder under BEAM repository, the initial simulation automatically generates network.dat and physsim.xml where all map-related data is stored.
  5. (households.csv, person.csv, plans.csv) The minimum requirements needed for conducting a BEAM run are households, persons, vehicle fleet, vehicle types, map data, and configuration. We assign virtual household and virtual driver for these truck trips.
  6. (linkAttributes_RIV.csv) To determine vehicle emission factors (in the unit of gram/mile), link-based traffic activities (e.g., average traffic speed) is needed  as an input. We get the linkspeed from BEAM Model Results and use them here.
  7. (facilityAttributes.csv, blockAttributes.csv) facility and residential block are prepared as the sensitive receptors.

Usage notes

LowExpRoutingMannual_v2.docx is the user manual for the Low exposure routing.

The minimum requirements are the fields in the uploaded files.


U.S. Department of Transportation, Award: 69A3551747109