The I-15 and I-80 Corridors in Salt Lake City Were Analyzed under Five Different Simulation Scenarios to Examine Impacts of Connected and Automated Vehicles (CAVs) at Crash Hotspots.
Salt Lake City, Ohio, United States
Impact of Connected Vehicle Technology on Traffic Safety under Different Highway Geometric Designs
Summary Information
Connected and Automated Vehicle (CAV) driving features, through their ability to increase awareness of the driving environment and improve driving behavior, have the potential to impact traffic safety in various ways. To mimic their behavior, smaller headways and shorter reaction times than human driven vehicles were modeled. This simulation study introduced the incorporation of CAVs as a new type of vehicle within the existing vehicle composition. The objective was to evaluate the impact of CAVs on traffic network safety, with a specific focus on the performance of traffic flow, including the examination of CAVs with distinct geometric designs and their potential to address crash hotspots. The model considered Salt Lake City in Utah as a case study. Five scenarios were simulated to address potential crash hotspots and the leading causes of crashes in these areas:
- Scenario-1 (I-15 northbound within the intersection of I-80 and 1300 South Street): The main objective was to study the lane changing behavior of CAVs in reduced lanes and their possible impacts on angle crashes.
- Scenario-2 (intersection of 100 South and University Street with higher traffic volume): This scenario was designed to illustrate a signalized intersection on a highly elevated layout.
- Scenario-3 (I-80 from Salt Lake City toward Parleys Canyon): This scenario had reduced visibility in cold weather seasons and provided variable speed limits (VSL).
- Scenario-4 (southbound I-15 freeway): Multiple consecutive on-ramps and off-ramps in this corridor were modeled.
- Scenario-5 (mountains, part of the I-80 freeway westbound from Parleys Canyon): The influence of a combination of horizontal and vertical alignments on road safety was modeled.
METHODOLOGY
A calibrated microsimulation model of I-210 was used to simulate approximately five minutes of traffic flow on the highway. Initially, the simulation was started with free-flowing traffic, then traffic was observed with flow smoothing-controlled vehicles inserted at various penetration rates. Vehicle models estimated fuel economy embedded within the simulation software. These models calculate the average MPG for simulated vehicles. Specifically, the energy consumption of electric vehicles was converted to an effective MPG by applying the National Resources Canada method and using the 2018 California electricity profile provided by the Energy Information Administration (EIA). The percentage of each vehicle model represented was determined by the 2018 registration data from the California Department of Motor Vehicles (DMV) and the Federal Highway Administration (FHWA), and 2019 sales data from the California New Car Dealers Association (CNCDA). Seven vehicle classes were used in the simulation, including three types of passenger vehicle classes, average light duty vehicle, average heavy-duty vehicle, passenger alternative fuel vehicle, and zero emission vehicle.
FINDINGS
- The average fuel economy achieved by the simulated human drivers increased from 19 MPG to 36 MPG, when the proportion of flow smoothing AVs increased from zero to 30 percent.
- The simulation results revealed a 50 percent increase in average speeds, when flow smoothing AV penetration rate increased from zero to 30 percent. This finding indicated a possible increase in vehicle miles traveled (VMT) by at most 30 percent.
- The average trip time in the simulations conducted decreased by 25 percent as the AV penetration rate increased from 0 to 30 percent.
- Among all of the vehicle types investigated, passenger alternative fuel vehicles and non-electric passenger vehicles improved the most in fuel economy, by about two times when flow smoothing AV penetration rate increased from zero to 30 percent. Electric vehicles, on the other hand, exhibited a less effect when implementing flow smoothing (about 1.2 times) which was likely owing to their regenerative braking that already resulted in significant energy savings.
