Robust Robotic Exploration and Mapping Using Generative Occupancy Map Synthesis

2025-06-24

AI Summary

To address incomplete and low-traversability 3D occupancy maps arising from partial observability in unknown environments, this paper introduces SceneSense—the first generative diffusion model for online robotic exploration that reconstructs 3D occupancy grids. SceneSense probabilistically infers occluded and unobserved regions from local sensor measurements and incrementally updates the global occupancy map via a Bayesian fusion mechanism. Fully compatible with standard motion planners, it integrates seamlessly into quadruped robot systems. Experiments demonstrate significant improvements over baselines: FID scores improve by 24.44% (near-range) and 75.59% (far-range); exploration trajectories exhibit higher consistency; and path planning success rate and robustness are markedly enhanced. The core contribution lies in introducing generative modeling to real-time occupancy mapping, enabling high-fidelity, traversability-aware 3D reconstruction under partial observability.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• 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: Economic ramifications for generative AI infrastructure and applications
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We present a novel approach for enhancing robotic exploration by using generative occupancy mapping. We introduce SceneSense, a diffusion model designed and trained for predicting 3D occupancy maps given partial observations. Our proposed approach probabilistically fuses these predictions into a running occupancy map in real-time, resulting in significant improvements in map quality and traversability. We implement SceneSense onboard a quadruped robot and validate its performance with real-world experiments to demonstrate the effectiveness of the model. In these experiments, we show that occupancy maps enhanced with SceneSense predictions better represent our fully observed ground truth data (24.44% FID improvement around the robot and 75.59% improvement at range). We additionally show that integrating SceneSense-enhanced maps into our robotic exploration stack as a "drop-in" map improvement, utilizing an existing off-the-shelf planner, results in improvements in robustness and traversability time. Finally we show results of full exploration evaluations with our proposed system in two dissimilar environments and find that locally enhanced maps provide more consistent exploration results than maps constructed only from direct sensor measurements.

Problem

• Enhancing robotic exploration using generative occupancy mapping
• Improving map quality and traversability with real-time predictions
• Validating performance in real-world experiments with a quadruped robot

Innovation

• Generative occupancy mapping enhances robotic exploration
• SceneSense predicts 3D maps from partial observations
• Real-time fusion improves map quality and traversability