TerrainForge: Physics-Grounded road geometry Editing for Counterfactual Autonomous Driving

2026-10-02

AI Summary

This study addresses the challenge of synchronously propagating road edits to vehicle dynamics, camera viewpoints, and inter-vehicle spacing in autonomous driving scenarios. It proposes the first framework that unifies the propagation of road geometry and surface condition changes to multi-vehicle dynamics, camera poses, and headway distances. The method constructs a unified road model linking scene deformation with four-wheel vehicle dynamics, generating physically consistent counterfactual driving videos from reconstructed multi-vehicle trajectory segments. Furthermore, a surrogate model is trained to enable low-cost safety screening. Validated through CarSim simulations and experiments on the Waymo dataset, the proposed approach reduces terminal spacing prediction errors by 22–40%, significantly enhancing both the efficiency and safety of counterfactual data generation.

Abstract

Road geometry (e.g., crests, sags, and speed humps) and surface conditions (e.g., wet or icy pavement) affect how vehicles move, what drivers and onboard cameras observe, and how much clearance remains between vehicles. Editing these properties in a driving scene therefore requires corresponding changes in vehicle motion. Capturing these differences in a driving video requires a road edit to propagate to vehicle motion, camera viewpoint, and the clearance between vehicles. We present TerrainForge, a framework for generating road geometry-focused counterfactuals from reconstructed multi-vehicle driving episodes. A unified road model connects scene deformation with four-wheel vehicle dynamics, allowing crests, sags, speed humps, and friction changes to propagate through vehicle motion, camera viewpoint, and inter-vehicle clearance. Vehicle dynamics are evaluated against CarSim, and prescribed road geometry is verified in reconstructed Waymo scenes. Across 18 episodes, leaving surrounding vehicles on their recorded trajectories instead of recomputing their responses produces median peak differences in predicted ego-lead distance of 1.52 m for crests and 1.41 m for sags. We further simulate the ego response to 15,758 road edits across 983 braking episodes, pairing each edit with its safety outcomes relative to an unedited replay. These pairs train a first-stage screening surrogate that takes the original driving context and candidate road-edit parameters as input and predicts the resulting change in the ego's terminal gap. On held-out scenes, this prediction achieves 22-40% lower mean absolute error than predicting no change, so candidates can be screened cheaply before the full multi-vehicle rollout.

Problem

• counterfactual driving
• road geometry editing
• vehicle dynamics
• autonomous driving
• scene generation

Innovation

• Counterfactual generation
• Vehicle dynamics
• Road geometry editing
• Surrogate model
• Autonomous driving