Terrain Costmap Generation via Scaled Preference Conditioning

2025-11-14

AI Summary

Terrain cost mapping for off-road navigation struggles to simultaneously achieve generalization across unseen terrain types and task-specific adaptability to user-defined cost preferences. Method: This paper proposes SPACER—a novel approach trained exclusively on synthetic data—that enables strong zero-shot generalization to previously unobserved terrain categories and supports real-time, user-driven adjustment of traversal cost preferences during inference. SPACER employs scaled preference conditioning for context-aware cost prediction and is validated on large-scale aerial terrain maps. Contribution/Results: Unlike existing semantic segmentation methods (which only permit rapid parameter tuning) or representation learning approaches (which only support generalization), SPACER achieves the lowest regret in 5 out of 7 real-world off-road path-planning benchmarks. It significantly enhances robotic navigation adaptability and robustness in complex, heterogeneous terrains.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Learning to Search
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Successful autonomous robot navigation in off-road domains requires the ability to generate high-quality terrain costmaps that are able to both generalize well over a wide variety of terrains and rapidly adapt relative costs at test time to meet mission-specific needs. Existing approaches for costmap generation allow for either rapid test-time adaptation of relative costs (e.g., semantic segmentation methods) or generalization to new terrain types (e.g., representation learning methods), but not both. In this work, we present scaled preference conditioned all-terrain costmap generation (SPACER), a novel approach for generating terrain costmaps that leverages synthetic data during training in order to generalize well to new terrains, and allows for rapid test-time adaptation of relative costs by conditioning on a user-specified scaled preference context. Using large-scale aerial maps, we provide empirical evidence that SPACER outperforms other approaches at generating costmaps for terrain navigation, with the lowest measured regret across varied preferences in five of seven environments for global path planning.

Problem

• Generate adaptable terrain costmaps for autonomous off-road navigation
• Enable rapid test-time cost adaptation while maintaining terrain generalization
• Overcome limitations of existing semantic and representation learning methods

Innovation

• Scaled preference conditioning for costmap adaptation
• Synthetic data training for terrain generalization
• User-specified context enables rapid test-time adaptation