Modeling the Lane-Change Reactions to Merging Vehicles for Highway On-Ramp Simulations

TL;DR

This paper addresses the challenge of accurately modeling how lag vehicles react to merging cars on highway ramps to improve high-fidelity AV simulations. It pairs the longitudinal MR-IDM model with two discretionary lane-change models, MOBIL and BRGT-D, and introduces a merge-adapted BRGT-D (mBRGT-D) to capture both keep-straight and lane-change behaviors. The authors construct HOMER, a new naturalistic merge dataset from eight U.S. ramps, and validate the models using exiD (German exits/entries) data, demonstrating real-time capable performance in a high-fidelity CarMaker-based simulation. Results show that behavior-specific parameter optimization yields higher predictive accuracy, with mBRGT-D achieving up to about $95\%$ lane-change prediction success compared to MOBIL's about $85\%$, indicating improved realism for merge interactions in simulation and providing a practical tool for AV development and testing.

Abstract

Enhancing simulation environments to replicate real-world driver behavior is essential for developing Autonomous Vehicle technology. While some previous works have studied the yielding reaction of lag vehicles in response to a merging car at highway on-ramps, the possible lane-change reaction of the lag car has not been widely studied. In this work we aim to improve the simulation of the highway merge scenario by including the lane-change reaction in addition to yielding behavior of main-lane lag vehicles, and we evaluate two different models for their ability to capture this reactive lane-change behavior. To tune the payoff functions of these models, a novel naturalistic dataset was collected on U.S. highways that provided several hours of merge-specific data to learn the lane change behavior of U.S. drivers. To make sure that we are collecting a representative set of different U.S. highway geometries in our data, we surveyed 50,000 U.S. highway on-ramps and then selected eight representative sites. The data were collected using roadside-mounted lidar sensors to capture various merge driver interactions. The models were demonstrated to be configurable for both keep-straight and lane-change behavior. The models were finally integrated into a high-fidelity simulation environment and confirmed to have adequate computation time efficiency for use in large-scale simulations to support autonomous vehicle development.

Figures (5)

• Figure 1: The distribution of national on-ramps and data collection sites selected to maximize range of speed category of on-ramp, speed category of highway, number of lanes on highway, acceleration lane length, and the on-ramp slope.
• Figure 2: Summary of the HOMER dataset collected across eight US sites in terms of total number of objects tracked, number of lane changes, and number of merges processed.
• Figure 3: The data collection setup (top) included two lidar sensors placed on the side of the road: a 128-beam primary lidar placed near the merging area (middle) and a 32-beam secondary lidar placed further down the road (bottom). Each lidar sensor was mounted on a base station mast, which included a camera platform, power supply, cellular WiFi router, data acquisition server, and monitor.
• Figure 4: Layout of relevant actors and their designations.
• Figure 5: Snapshot of simulation in CarMaker. Actor T02 is performing a lane change in response to the merging vehicle.